INTRODUCTION
Background
The Vietnamese stock market, established in 1998 with the Ho Chi Minh City Stock Exchange (HOSE) and Hanoi Stock Exchange (HNX), officially launched on July 28, 2000, featuring only two listed companies and four securities firms Over the next 14 years, the number of securities companies grew to 181, yet the market remains relatively small compared to developed and emerging markets Despite significant growth in listed stocks and transaction values on HOSE, price fluctuations have been unpredictable The VN-Index surged from 100 points in July 2000 to a peak of 571 points by June 2001, driven by investor excitement and rising demand amid limited listings However, a lack of investor knowledge and insufficient regulatory support led to a dramatic decline, with the VN-Index plummeting to 139 points by March 2003 This downturn left many investors facing substantial losses, causing the market to stagnate until its revival in 2006, when a new boom began in the second half of the year.
As of May 20, 2014, data reveals that the Ho Chi Minh City and Ha Noi stock markets experienced a significant surge, reaching 1,170 points by March 2007 Following this peak, the index fluctuated around 1,000 points until October 2007, marking a notable period in the performance of the Ho Chi Minh City stock market.
The VN-Index experienced a significant decline after reaching its peak, marking a challenging year for the Ho Chi Minh City stock market in 2008 It wasn't until February 2009 that the index stabilized at 235 points (Huy, 2010) Several factors contributed to this sharp drop, including tightened monetary policies, elevated deposit interest rates, high inflation, and the recession of the U.S economy Additionally, the lack of timely intervention by authorities exacerbated the decline of the VN-Index (Vo and Pham).
2008) From 2009 to the first quarter of 2011, VN-Index continued to undergo many ups-and-downs, reached another peak at 542 points in May
2010 and another bottom at 351 points in November 2011 Nonetheless, it seemed to fluctuate between 400 and 500 points with no significant amplitude found in 2012 and finally stood at 505 points on December 31 st
In 2013, stock prices experienced a significant increase of 23% compared to 2012 (Luyen, 2013) According to Phu (2010), herding bias plays a crucial role in stock price fluctuations When stock prices rise, many investors buy, anticipating further increases, but prices often peak and then decline Conversely, when prices fall, fear drives investors to sell, creating high selling pressure This behavior indicates that investors often act based on emotions rather than rational analysis, disrupting the market's supply-demand dynamics and potentially leading to market collapse Therefore, herding bias is a critical factor that warrants in-depth analysis, particularly in the context of the Vietnamese stock market.
Expected Utility Theory (EUT) serves as the foundational principle of traditional finance, while prospect theory underpins behavioral finance, emphasizing the impact of investors' subjective value systems on decision-making EUT, recognized as both a normative model of rational choices and a descriptive model of economic behavior, has been pivotal in analyzing decisions under risk However, it faces criticism for not adequately explaining human attraction to both insurance and gambling Individuals tend to undervalue probable outcomes in favor of certain ones, responding differently to similar situations based on the context of losses and gains Prospect theory highlights psychological factors influencing decision-making, such as regret aversion, loss aversion, and mental accounting.
Based on the background, the problem statement would state my concern about the impact of behavioral factors on herding bias in the case of Vietnamese stock market.
Problem statement
Established in 2000, the Vietnamese stock market is marked by weak reporting requirements, inadequate regulations, and low accounting standards (Tran and Truong, 2011) Throughout its development, the market, especially the HOSE, has experienced various stages, leading to unpredictable price fluctuations that hinder rational decision-making for investors Despite these dramatic price changes and overall market instability, there is a notable lack of research focused on the Vietnamese stock exchange, particularly regarding herding behavior among investors.
Conventional financial theories have struggled to elucidate the recent developments at the HOSE, making behavioral finance a more effective framework for understanding market dynamics This approach leverages psychological factors to explain stock trading behaviors, as highlighted by Waweru et al (2008) Key behavioral influences include overconfidence, representativeness, availability, loss aversion, regret aversion, gambler's fallacy, overreaction and underreaction, and herding tendencies, as noted by Ritter.
Herding behavior among investors is a significant factor contributing to excess volatility and short-term trends in financial markets (Juan Yao et al., 2014) Research conducted on the Vietnamese stock market, including studies by Farber et al (2006) and Tran (2007), has shown a pronounced herding effect, particularly in response to positive market returns Farber et al (2006) highlighted the influence of policy on the Vietnamese stock market and provided empirical evidence of herding behavior, indicating that investors often disregard their private information and expectations to align with the collective actions of the market.
“the trend of herding behavior is stronger toward extreme positive returns of the market, and in fact, around the consecutive sequence of limit-hits” (Farber et al, 2006, p.25)
Chen et al (2003) suggested that herding behavior is more prevalent in emerging markets compared to developed markets due to higher government intervention and lower quality of information disclosure This government involvement disrupts the market's natural functioning, leading to a less efficient allocation of resources.
“demand-supply” rule; therefore, there was no random trend in the market
In emerging markets, the low quality of information disclosure leads to a reliance on rumors and unofficial sources, contributing to market instability Kaminsky and Schmukler (1999) noted that during the 1997-1998 Asian financial crisis, herding behavior was prevalent, particularly in Vietnam Tran and Truong (2011) highlighted that Vietnamese investors often mimic the actions of those perceived to be better informed, a tendency exacerbated by the regulatory framework's inadequacies, such as a lack of transparency and efficient reporting mechanisms This informational inefficiency, coupled with high market volatility, drives investors to make consensus-based decisions, resulting in increased correlations among stock returns and a reduction in return dispersion Consequently, herding behavior is evident in the Vietnamese stock market, prompting this study to analyze the herding effect and its influencing factors.
Understanding the herding bias in the Vietnamese stock market requires an exploration of the behavioral factors influencing this phenomenon at the HOSE By identifying these factors, investors can gain insights into common investment behaviors, enabling them to make informed decisions for improved returns Additionally, security companies can leverage this information to better understand investor behavior, leading to more accurate forecasts and enhanced recommendations Ultimately, this understanding will help ensure that stock prices reflect their true value, positioning the HOSE as a benchmark for economic prosperity and facilitating capital raising for corporate growth and expansion.
Research question
This research investigates the behavioral factors influencing herding bias in the Ho Chi Minh Stock Exchange (HOSE) since its establishment in 2000 It aims to assess the impact of these behavioral elements on herding bias among investors Consequently, the study poses the research question: What behavioral factors contribute to investors' herding bias in the Vietnamese stock market?
Research scope
The Vietnamese stock market consists of two primary exchanges: the Ho Chi Minh City Stock Exchange (HOSE) and the Hanoi Stock Exchange (HNX) This study will focus specifically on the HOSE for for analysis.
Ho Chi Minh City stands as Vietnam's largest and most rapidly developing urban center The Ho Chi Minh Stock Exchange (HOSE) serves as a key indicator of economic wealth and a vital platform for businesses to secure funding for growth and production (Luong and Ha, 2011) Choosing the HOSE for this study facilitates easier data collection, improves the accuracy of survey responses, and ensures more dependable interview results, all while adhering to budget and time limitations.
Research methods
This research will utilize quantitative methods, specifically incorporating a pilot study and a survey The quantitative analysis will involve distributing survey forms to over 200 investors participating in the HOSE market.
This study aims to develop hypotheses grounded in established behavioral finance theories, utilizing a quantitative method for initial testing To ensure the reliability of measurement scales, the research will employ SPSS software to calculate Cronbach’s Alpha using standardized items Following this, the validity of the measurement scales will be assessed through Exploratory Factor Analysis (EFA) Ultimately, the paper will implement regression analysis to evaluate the proposed model and test the hypotheses.
Significance of the research
This research has several meanings for individual investors, security organizations, the field of behavioral finance as well as for the author
Individual investors can benefit from research as it serves as a valuable reference for understanding stock investment behavior By analyzing market trends and recognizing herding bias, investors can make informed and suitable investment decisions.
The research offers security companies valuable insights into the behavioral factors influencing herding bias, enabling them to enhance the accuracy of their forecasts and provide more reliable advice to investors.
Behavioral finance, a relatively new field compared to traditional financial theories, has been widely applied in developed stock markets but remains underutilized in less developed markets like Vietnam Research on herding behavior in emerging markets is limited, highlighting a gap in understanding This study aims to demonstrate the applicability of behavioral finance across all stock market types and contribute valuable insights into the herding effect in Vietnam's financial landscape.
Structure of the study
The proposal of this study has 3 chapters:
This chapter is divided into two sections: the first provides a historical overview of the HOSE from its inception to the current state, while the second outlines the problem statement and explains the rationale behind selecting this topic for discussion.
This chapter reviews behavioral factors influencing herding bias as identified in prior studies, focusing on elements such as representativeness, monitoring, and anchoring It proposes an examination of five key behavioral factors: risk tolerance, overconfidence, self-monitoring, availability bias, and gambler’s fallacy Additionally, this section will outline the development of related hypotheses.
This chapter outlines the quantitative research method utilizing a survey approach, where I will administer questionnaires to over 200 market participants The questionnaires include items designed to measure relevant variables Subsequently, I will assess the scale measurements and apply regression analysis to evaluate the proposed model and hypotheses.
Chapter 4: Data analysis and finding
This chapter presents descriptive statistics and refines measurement scales, highlighting the results of Cronbach's alpha and factor analysis from the formal survey.
Furthermore, the results of testing the assumption of regression and hypothesis testing are also mentioned
At the end of the chapter, a general summary would be stated
Throughout this chapter, I would present a summary of main findings, discuss the meaning and contribution in investment practice
To finish the thesis, a limited number of topics and further research direction would be also stated
LITERATURE REVIEW
Theorical background
The Efficient Market Hypothesis (EMH) and the Random Walk Model (RWM) have dominated financial literature for the past three decades Fama (1970) introduced the EMH, suggesting that investors trust current market prices, which reflect all available information about a security In parallel, the RWM posits that past price movements do not provide opportunities for excess returns While not all investors act rationally, the EMH asserts that markets are efficient and capable of making unbiased predictions.
In 1970, the Efficient Market Hypothesis (EMH) was categorized into three forms based on the availability of information: weak, semi-strong, and strong The weak form asserts that current asset prices reflect all historical financial data available in the market The semi-strong form indicates that share prices quickly and unbiasedly adjust to new publicly available information, meaning no excess returns can be earned from trading based on that information Finally, the strong form encompasses both weak and semi-strong efficiencies, suggesting a comprehensive market efficiency framework.
To be more exact, it implies that share prices reflect all kind of information which may be public and private, but none of these can earn excess returns
Behavioral finance challenges the notion of efficient markets by suggesting that they can be influenced by psychological factors (Ritter, 2003) Tversky and Kahneman (1979), pioneers in this field, elucidate how human behavior affects financial decisions at various stages Olsen (1998) describes behavioral finance as a new paradigm aimed at enhancing the traditional finance theories, while Fromlet and Hubert (2001) emphasize its integration of individual behaviors and market dynamics, drawing from psychology and financial theory Furthermore, Olsen (1998) highlights that behavioral finance focuses on applying psychological and economic principles to improve financial decision-making This discipline enriches our understanding of individual investors by revealing the psychological traits and processes that shape their investment intentions and choices (Ritter, 2003).
Prospect theory is a fundamental concept in behavioral finance, highlighting how psychological factors influence decision-making Key components of this theory include regret aversion, loss aversion, and mental accounting, which collectively shape individuals' financial choices and perceptions of risk (Waweru et al., 2003, p.28).
Prospect theory suggests that regret is an emotion experienced after making mistakes, leading investors to avoid selling shares that are decreasing in value while being more willing to sell those that are increasing (Forgel & Berry, 2006) Loss aversion highlights the greater mental penalty individuals feel from losses compared to equivalent gains (Barberis & Huang, 2001), with studies indicating that the distress from potential losses outweighs the pleasure derived from similar gains (Barberis and Thaler, 2003) Additionally, mental accounting refers to how individuals evaluate their financial transactions, allowing investors to organize their portfolios into distinct accounts (Barberis & Thaler, 2003).
In summary, the EMH means that there is unsystematic method to beat the market and stock prices are reasonable, namely, they only reflect
Behavioral finance focuses on the psychological aspects of decision-making in financial markets, contrasting with traditional views that emphasize fundamental characteristics like risk It acknowledges the impact of psychological factors, such as sentiment, on investment choices The introduction of concepts like prospect theory marks a significant shift in financial theory, challenging the Efficient Market Hypothesis (EMH) by underscoring the importance of behavioral influences in predicting market movements.
Review on some behavioral factors and herding bias in stock market
Numerous researchers have examined the behavioral factors influencing herding bias, including risk tolerance, overconfidence, gambler's fallacy, illusion of control, and self-monitoring This section will present a literature review of these critical factors.
Financial risk tolerance refers to the level of uncertainty an individual is willing to accept in financial decision-making and is influenced by various demographic and socioeconomic factors, such as gender, age, marital status, ethnicity, and income Research indicates that as age increases, risk tolerance tends to decrease, while higher income levels are associated with increased risk tolerance Studies show that single individuals generally exhibit greater risk tolerance compared to their married counterparts, with single males demonstrating the highest levels, followed by married males, unmarried females, and married females Additionally, findings suggest that men typically possess a higher risk tolerance than women.
Other factors that appear to impact a person’s risk tolerance include environmental factors such as financial knowledge, family situation (Roszkowski, 1999) and social development such as birth order (Sulloway,
1997), which is an example of a biopsychosocial factor
Research indicates that financial risk tolerance is a significant positive predictor of stock investment willingness, with highly risk-tolerant investors maintaining high-value portfolios and engaging in frequent trading (Keller & Siergist, 2006) Studies by Dorn and Huberman (2005) further reveal that these investors exhibit more aggressive trading behaviors, leading to increased portfolio diversification and turnover Multiple studies consistently show that individuals with higher risk tolerance tend to trade more frequently compared to their less risk-tolerant counterparts (Tigges et al., 2000; Warneryd, 2001; Clark-Murphy & Soutar, 2004; Wood & Zaichkowsky, 2004; Durand et al., 2008).
Previous research indicates that attitudes toward financial risk tolerance significantly influence the setting of financial goals and the formulation of financial plans and strategies (Grable and Joo, 2004), potentially affecting herding behavior in financial decision-making.
Herding bias can be influenced by the differences in risk tolerance and risk preferences among investors, as suggested by Lin (2012) He posited that lower risk tolerance significantly contributes to the prevalence of herding behavior in financial markets.
Numerous studies have explored the phenomenon of overconfidence in investors Kourtidis et al (2011) highlight that overconfident investors tend to be overly certain of their abilities, often neglecting the perspectives of others This overconfidence leads them to under-react to new information and to overvalue certain data, resulting in unrealistic expectations regarding their potential returns (Barber and Odean).
Overconfident investors tend to overestimate their private information while disregarding available data, leading to distinct market behaviors such as short-horizon momentum and long-horizon reversal in returns (Daniel et al., 1998) This overconfidence is associated with increased trading frequency and volume, as noted by Kourtidis et al (2011), who found that a higher degree of overconfidence correlates with elevated trading activity (Glaser and Weber, 2007) Additionally, Dow and Gorton (1997) observed that trading volume rises when both individuals and insiders exhibit overconfidence, further supporting the notion that overconfidence drives greater trading activity.
(2001), Hirshleifer and Luo (2001), Wang (2001) and Scheinkman and Xiong
Research indicates that overconfidence significantly increases trading activity and the likelihood of making poor decisions, such as purchasing the wrong stocks (Kourtidis et al., 2011) Odean (1998) found that overconfident traders often exhibit biased judgments that can result in reduced returns Additionally, studies by Fenton O’Creevy et al (2003) and Philip (2007) confirm that overconfidence negatively impacts trading performance Conversely, De Long et al (1990) and Wang also contribute to the understanding of these dynamics in trading behavior.
(2001) supported that overconfident investors earned higher returns than less confident ones
The relationship between overconfidence and herding behavior indicates that when investors exhibit overconfidence, their tendency to follow the crowd diminishes Research on overconfident trading behaviors reveals that individual investors often overestimate their information and abilities, as shown in studies by Barber and Odean (2001a, 2001b).
Research indicates that certain investors, influenced by behavioral biases, often underestimate risk and engage more frequently in high-risk securities This tendency results in investment performance that frequently falls below the market average.
Self-monitoring is an important personality trait in social psychology, often viewed as a form of social intelligence It involves the ability to observe social cues and adapt one's behavior to meet the expectations of a given social environment (Snyder and Gangestad, 1986) According to Parker and Fischhoff (2001), this trait has garnered increased attention for its role in social interactions.
Decision-making competence is positively correlated with self-monitoring, which reflects an individual's awareness of their actions and has been linked to enhanced performance Research by Kilduff and Day (1994) indicates that high self-monitors are more likely to achieve promotions in managerial roles, while Mehra, Kilduff, and Brass (2001) found that self-monitoring positively impacts workplace performance High self-monitors, often seen as impression managers, strategically adjust their behaviors to create advantageous impressions in various situations In trading contexts, this translates to strategic and manipulative behaviors, such as placing orders to profit without disclosing private information to other participants They may also make offers that do not align with their true beliefs, aiming to influence others' perceptions (Biais et al., 2005) Additionally, Monson (1983) identified a projection effect where high self-monitors assume others behave similarly, interpreting others' actions as influenced by situational factors rather than inherent traits Consequently, in market scenarios, they are less likely to take prices at face value and are more attuned to the strategies behind market signals, reducing the likelihood of falling victim to the winner’s curse (Eyster and Rabin, 2003).
The gambler’s fallacy refers to the mistaken belief that the probability of an event occurring decreases after it has already happened, despite the independence of each event (Rabin, 2002) Essentially, gamblers assume that small samples should reflect the overall population, leading them to expect correlations when unexpected sequences arise (Tversky and Kahneman, 1971) In the context of stock markets, this fallacy manifests when investors misjudge reversal points, believing they signal the end of favorable or unfavorable market trends (Waweru et al., 2008) Furthermore, when influenced by this bias, investors often make suboptimal choices based on previous selections (Kempf and Ruenzi, 2006).
The illusion of control, as described by David and Meira (1995), is the belief that individuals can influence the outcomes of random events This phenomenon arises from a difficulty in distinguishing between skill-based situations, where outcomes are influenced by one's actions, and chance events, which are purely random When people try to apply skillful actions to affect outcomes in situations governed by chance, they are exhibiting an illusion of control.
The illusion of control manifests through various dependent measures, including willingness to trade lotteries, amounts wagered on uncertain outcomes, selling prices, and confidence ratings Research by Langer (1975, 1977) indicated that this illusion can be amplified by manipulating factors that participants associate with skill, such as competition, choice, familiarity with the stimulus or response, and levels of involvement, whether active or passive.
Hypothesis development
Prospect Theory is a fundamental to explain the fluctuation of stock market
To be more specific, prospect theory describes some states of mine affecting an individual’s decision making process including Regret aversion, Loss aversion and Mental accounting (Waweru et al, 2003, p.28)
In accordance with the research of Ly and Thao (2012), over-confidence, risk tolerance, and herding bias have significant effects on Vietnamese investors
This research explores the behavioral factors influencing herding bias among investors in the Vietnam stock market It highlights that many Vietnamese investors base their buying and selling decisions on recent information rather than fundamental analysis, suggesting that availability bias plays a significant role in herding behavior Additionally, existing studies indicate that self-monitoring and gambler's fallacy significantly impact biases in stock markets Consequently, the research hypothesizes that four key behavioral factors—risk tolerance, overconfidence, self-monitoring, and gambler's fallacy, along with availability bias—affect herding bias in this context.
Yang and Qiu (2001) identified that investors who exhibit anxiety and emotional instability tend to follow the investment advice of friends or seek professional guidance, which can lead to herding behavior In contrast, high risk-tolerant investors, who are more accepting of uncertainty, are less likely to engage in herding Garble and Lytton (2004) further suggested that understanding the link between stock market returns and risk tolerance may clarify why investors display herding by buying risky assets during market uptrends and selling during downturns Lin (2012) posited that the root cause of herding behavior could stem from differences in risk tolerance and preferences, leading to the conclusion that risk tolerance significantly negatively impacts herding bias Thus, the first hypothesis is established.
H1: Risk tolerance has a negatively effect on herding bias
Individual investors often exhibit herding behavior due to a lack of confidence and professional knowledge, which leads them to rely on market signals and the opinions of professional investors when making investment decisions.
Overconfidence significantly influences herding bias in investors, as noted by Goodfellow, Bohl, and Gebka (2009) When investors exhibit high levels of confidence, they tend to rely more on their private information for making investment decisions Consequently, this reliance diminishes their interest in following the herd behavior in the market.
H2: Overconfidence has a negative effect on herding bias
Self-monitoring plays a significant role in influencing herding bias among market participants Individuals with high self-monitoring tendencies are adept at predicting how their actions affect others and anticipate that others will behave strategically as well (Biais et al., 2005) This awareness leads them to be skeptical of market prices, prompting them to analyze the underlying signals and strategies that drive these prices rather than accepting them at face value As a result, high self-monitors exhibit reduced interest in herding behaviors, supporting the proposed hypothesis 3.
H3:Self-monitoring has a negative effect on herding bias
Gambler's fallacy significantly influences herding bias, as individuals affected by this fallacy mistakenly believe that the likelihood of an event occurring decreases after it has already happened, despite the independence of each trial This belief often leads them to trade contrary to recent trends, ultimately inhibiting herding behavior (Rabin and Vayanos, 2009) Therefore, we propose the following hypothesis:
H4: Gambler’s fallacy has a negative effect on herding bias
The illusion of control is a key factor influencing herding bias, where individuals believe they can manipulate random events and mistakenly view chance games as skill-based (Langer, 1975; Kahneman and Riepe, 1998) This bias leads investors to draw conclusions from the actions of others, assuming those decisions are informed rather than random As a result, the illusion of control fosters herding behavior among investors.
H5: Illusion of control has a positive effect on herding bias
As is presented above, the research model is organized as follows
Year of publication Antecedent of hearding bias Sign of effect
Over-confidence, risk tolerance, and herding bias have significant effects on Vietnamese investors
The behavioral factors of investors affecting the herding bias in the Vietnam stock market
Self-monitoring and gambler’s fallacy have considerable effects to biases in stock markets
The research hypothesises that the four behavioral factors affecting herding bias are risk tolerance, overconfidence, self- monitoring, gambler’s fallacy, and availability bias
Investors having characteristics of anxiety, emotionally unstable and nervous always followed the investment suggestions of their friends or seek professional consultation or insider that will also lead to herding
Investors having characteristics of anxiety, emotionally unstable and nervous always followed the investment suggestions of their friends or seek professional consultation or insider that will also lead to herding
Understanding the relationship between stock market returns and risk tolerance
Explaining why investors exhibited herding behavior by purchasing risky investments during market uptrend and selling securities during market downtrend
Risk tolerance had a significant negative impact on herding bias
The original herding behavior might be the differences of risk tolerance and risk preferences of the behavior
A chart of research model proposed by the author
RESEARCH METHODOLOGY
Measurement scales
3.1.1 Scales measurement of Risk Tolerance
Variable of Risk Tolerance is symbolized of RIS Trone, Allbright and Taylor
Measuring an individual's financial risk tolerance is challenging due to its multidimensional nature and the various predisposing factors that influence it (1996) In this study, Risk Tolerance (RIS) will be assessed using a five-point Likert scale, as proposed by Grable and Joo (2004) This variable will be evaluated through five specific items, as outlined in the accompanying table.
Many individuals perceive investing as a complex endeavor, often preferring the security of a bank account over the uncertainties of the stock market The association of "risk" with "loss" can deter potential investors, leading them to believe that success in stocks and bonds relies heavily on luck Consequently, the emphasis on safety often outweighs the pursuit of higher returns in their investment decisions.
3.1.2 Scales measurement of Over-confidence
The variable of over-confidence, referred to as CONFI, will be assessed in this study using five-point Likert scale items derived from the research conducted by Lin (2011) and Lakshmi et al (2013) The accompanying table outlines these measurement items.
Confi1 I am sure that I can make the correct investment decision
Confi2 I believe I can master the future trend for my investment
I believe that market trends align well with my perspectives, and I consistently attribute my investment profits to my successful strategies My previous profitable investments can be largely credited to my specialized investment skills Additionally, the return rate of my investments meets or exceeds the average market return, leaving me satisfied with my past investment decisions.
3.1.3 Scale measurement of Self-monitoring
Self-monitoring, represented by the acronym SEMO, is a stable personality trait that persists throughout an individual's life, as evidenced by Jenkins (1993) Snyder and Gangestad (1986) created and validated a psychometric scale to measure self-monitoring This paper will include their 18-item questionnaire, which is presented in the following table.
Semo1 I find it hard to imitate the behavior of other people
Semo2 At parties and social gatherings, I do not attempt to do or say things that others would like Semo3 I can only argue for ideas which I already believe
Semo4 I can make impromptu speeches even on topics about which I have almost no information Semo5 I guess I put on a show to impress or entertain others
Semo6 I would probably make a good actor
Semo7 In a group of people I am rarely the centre of attention
Semo8 In different situations and with different people, I often act like very different persons Semo9 I am not particularly good at making other people like me
Semo10 I'm not always the person I appear to be
Semo11 I would not change my opinions (or the way I do things) in order to please someone or win their favor Semo12 I have considered being an entertainer
Semo13 I have never been good at games like charades or improvisations
Semo14 I have trouble changing my behavior to suit different people and different people and different situations Semo15 At a party I let others keep the jokes and stories going
Semo16 I feel a bit awkward in public and do not show up quite as well as I should Semo17 I can look anyone in the eyes and tell a lie with a straight face
Semo18 I may deceive people by being friendly when I really dislike them
3.1.4 Scale measurement of Gambler’s Fallacy
The gambler's fallacy variable, denoted as GAM, is measured using a four-item questionnaire based on a five-point Likert scale, as outlined by Waweru et al (2008) The items designed to assess the gambler's fallacy are detailed below.
GAM Gambler's Fallacy gam1 You guess reserve points to decide your investment gam2
Price of a stock has fallen in multiple sessions You determine to buy this stock because you believe that it would be impossible to decline more gam3
The stock price has risen consistently over several sessions, prompting you to decide to sell your shares, as you doubt its potential for further gains Regardless of this, you remain committed to investing in stocks based on thorough fundamental or technical analysis.
3.1.5 Scale measurement of Illusion of control
“Illusion of control” variable is symbolized as ILLUS According to Scottand Mariathis, variable could be measured by eleven five-point-Likert items The following table presents these items
Many gamblers fall into the illusion of control, believing that a systematic approach to playing slot machines or other games can lead to greater success For instance, a common misconception is that after a coin has landed heads ten times in a row, the next toss is more likely to be tails However, some individuals believe there are secrets to successful casino gambling that can be learned, while others attribute their wins to being "born lucky." Gamblers often think that wins are more likely on "hot" machines, and a skilled player is akin to a quarterback who knows the right strategies to employ at the right time It’s often advised to stick with the same pair of dice during a winning streak, as a well-planned system can lead to consistent victories Additionally, some players think that the longer they've been losing, the more likely they are to win, and familiarity with a game can increase their chances of success Lastly, paying attention to frequently winning lottery numbers can also be a part of a strategic approach to gambling.
The herding bias variable, denoted as HERD, will be assessed using a total of eight five-point Likert items derived from two prior studies Specifically, four items will be sourced from Lin (2011) and four from Lakshmi et al (2013) The complete set of these eight-item questionnaires is presented in the following table.
Herding bias in investing often leads individuals to make decisions based on the actions and recommendations of others rather than their own analysis For instance, many investors tend to buy shares based on friends' suggestions, follow rising stock prices, or mimic the trading choices of peers Additionally, they may rely on market information, technical analysis, media coverage, and expert opinions, as well as insights from family and friends, to guide their investment strategies This behavior highlights the influence of social dynamics on financial decision-making.
Sampling
Hatcher (1994) states that for effective factor analysis, the sample size must be at least five times the number of observed variables or a minimum of 100 participants In this study, with 37 observed variables, the recommended sample size is N = 5 x 37 = 185 This research will include over 200 participants, meeting the necessary requirements Additionally, Tabachnick and Fidell (2011) emphasize that for regression analysis, the sample size should be equal to or greater than a specified threshold.
50 + 8*k (N>P + 8*k), in which k is the number of independent variables In this study, N > 50 + 8*5 Hence, the survey form will be sent to more than
200 investors in each stock exchange all met the given condition.
Data collection methods
This study will utilize a quantitative approach, focusing on a targeted survey of over 200 investors on the HOSE Initially, selective questionnaires will be distributed to gauge the level of agreement among participants.
A pilot survey involving 20 investors will be conducted to assess the reliability of measurement scales The reliability test will utilize SPSS software to calculate Cronbach’s Alpha based on standardized items After removing unreliable questions, the refined questionnaires will be distributed to the remaining investors This process will continue until a sufficient number of responses is collected.
A total of 200 questionnaires from investors were completed to evaluate the scales' reliability and validity through exploratory factor analysis (EFA) with a sample size of N=5 The final step involves utilizing SPSS software to conduct regression analysis to test four key assumptions.
Data analysis methods
3.4.1 Test of scale measurement reliability
The Cronbach’s Alpha Test is utilized to assess the reliability of scale measurements using a five-point Likert scale, ensuring consistency among respondents' answers This test not only evaluates the reliability of the measurement but also aids in predicting the dependability of respondents' feedback (Helm, Henze, Sass, and Mifsud, 2006) As noted by Liu, Wu, and Jumbo (2010), Cronbach’s Alpha is a common indicator of reliability in behavioral research, making it particularly appropriate for studies involving behavioral finance that utilize a five-point Likert questionnaire.
As a consequence, this paper will employ Cronbach’s Alpha to test of scale measurements reliability including the factors built after the factor analysis
Cronbach’s Alpha is a crucial measure of reliability in research, with Nunnally (1978) suggesting a minimum value of 0.7 for acceptable measurements, while some statisticians consider a value over 0.6 to be sufficient It is essential to evaluate corrected item-total correlations, which should ideally be 0.3 or higher (Shelby, 2011) In this study, the acceptable Cronbach’s Alpha range is set between 0.7 and 0.8, and the corrected item-total correlation must meet or exceed 0.3, particularly as the financial behavior measurements are novel to stockholders on the Ho Chi Minh Stock Exchange Furthermore, the significant level for the F-test in the Cronbach’s Alpha analysis should not exceed 0.05, and the testing will be conducted using SPSS software.
This study utilizes Exploratory Factor Analysis (EFA) to identify the underlying factors related to the behavioral finance variables from questions 1 to 37 of the questionnaire EFA serves to eliminate items that do not meet the analytical criteria, as outlined by O’Brien (2007), and to test the hypotheses proposed in the research model presented in Chapter 3 The criteria for the EFA conducted in this research include factor loadings, the Kaiser-Meyer-Olkin (KMO) measure, total variance explained, and Eigen-values.
Factor loadings represent the correlations between individual items and their respective factors According to Hair et al (1998), for exploratory factor analysis (EFA) to yield practically significant results, the factor loadings should exceed 0.5 when the sample size is 100.
The Kaiser-Meyer Olkin Measure of Sampling Adequacy (KMO) reveals the level of appropriation of using EFA for the collected data Ali, Zairi and Mahat
(2006) suggested that the KMO should be flowed from 0.5 to 1.0(significant level less than 0.005) to ensure that factor analysis was suitable for the data
To determine the number of retained factors, total variance explained will be analyzed, ensuring that only factors contributing a significant amount to the explained variance are included According to Hair et al (1998), it is recommended that the total variance explained exceeds 50% to ensure meaningful results.
Eigen-value represents the variance explained by a factor in a dataset, with a value greater than 1 indicating that the factor explains more variance than a single item Values below 1 suggest that the factor contributes less information than individual variables (Leech, Barret, and Morgan, 2005) Both Eigen-value analysis and Cronbach’s Alpha calculations are performed using SPSS software.
To explore the correlation between herding bias and other behavioral factors,
I will also apply the regression analysis The model is presented as follow: HERD In which:
HERD stands for Herding bias
ILLUS: Illusion of control is constant
DATA ANALYSIS AND FINDINGS
Descriptive statistics
A total number of 205 questionnaires were delivered to investors in HOSE
The final sample consisted of 205 participants, and data analysis was conducted using SPSS software The assessment of normality revealed that the significance values for all variables were greater than 05, a typical occurrence in large samples Additionally, the 5% trimmed mean values closely aligned with the original mean values, leading to the decision to retain all outliers in the dataset.
In this study, variables are assessed using a multi-item scale, specifically a 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree) The average values of the observed variables are calculated to represent the concepts essential for the research, with descriptive statistics summarized in Table 4.1.
Table 4.1: Independent variables, dependent variables and items
1 Risk Tolerance RIS1, RIS2, RIS3, RIS4, RIS5
2 Over-confidence CONFI1, CONFI2, CONFI3, CONFI4, CONFI5, CONFI6,
SEMO1, SEMO2, SEMO3, SEMO4, SEMO5, SEMO6, SEMO7, SEMO8, SEMO9, SEMO10, SEMO11, SEMO12, SEMO13, SEMO14, SEMO15, SEMO16, SEMO17, SEMO18
4 Gambler's Fallacy GAM1, GAM2, GAM3, GAM4
5 Illusion of control ILLUS1, ILLUS2, ILLUS3, ILLUS4, ILLUS5, ILLUS6,
ILLUS7, ILLUS8, ILLUS9, ILLUS10, ILLUS11
Herding bias HERD1, HERD2, HERD3, HERD4, HERD5, HERD6,
Refinement of measurement scales
4.2.1 Result of Cronbach’s alpha analysis of formal survey (N 5)
Following the formal survey, certain questions were discarded due to their Cronbach's Alpha scores falling below the acceptable threshold of 0.6 The remaining behavioral factor questions, ranging from X6 to X50, demonstrated reliability as independent variables Meanwhile, questions Y51 to Y59, categorized as dependent variables, were designed to assess investors' evaluations of their own herding behavior The Cronbach’s Alpha scores for the independent variables are detailed below.
Table 4.2 Results of Cronbach’s alpha for behavioral factors
The behavioral factors Cronbach’s alpha Benchmark
Table 4.2 reveals that the Cronbach’s alpha values for Risk Tolerance (0.658), Over-confidence (0.61), Self-monitoring (0.698), Gambler’s Fallacy (0.699), Illusion of Control (0.766), and Herding Bias (0.795) are all above the acceptable threshold of 0.6, indicating satisfactory reliability for these constructs Detailed analysis of these Cronbach’s alpha results can be found in Table 4.2.
Table 4.3 indicates that the corrected items-total correlation for all variables exceeds 0.3, demonstrating sufficient reliability for each variable, which were subsequently utilized in the Exploratory Factor Analysis (EFA) Detailed Cronbach's alpha results for all items, analyzed using SPSS, are provided in Appendix 2.
Exploratory factor analysis (EFA) was conducted on the behavioral variables related to herding behavior, beginning with question X6, to identify the factors influencing this behavior The analysis met the necessary requirements outlined in Chapter 3, allowing for the effective reduction of items.
After conducting several rounds of item removal, the analysis revealed that the remaining data was categorized into five variables, comprising four independent variables and one dependent variable This research involved eliminating several items that failed to meet the established criteria As outlined in Chapter 3, it is essential for the KMO value to exceed the minimum threshold for valid results.
0.5 Factor loading is greater or equal 0.5, i.e if factor loadings of the items were less than 0.5, they would be got rid of The total variance explained was suggested to be more than 50% In terms of absolute values of difference among loading, the acceptable value was at least 0.3 The Corrected Item-Total correlation should be at least 0.4 to ensure that the items would make a good component of a summated rating scale Eigen-value should be greater than 1 because Eigen-value less than 1 means that less information was explained by the factor than by a single item (Leech, Barrett & Morgan, 2005, p.82) a Individual
Table 4.3 presented the brief results of EFA As is shown, KMO values of all factors are great or equal 0.5, therefore, factor loading was suitable for data of survey
Table 4.3 KMO index of behavioral factors
The behavioral factors KMO Benchmark
Appendix 3 shows that factor loadings are greater than 0.5, except for RIS5 (0.487) The values of the other indicators including total variance explained, Corrected Item-Total correlation, Eigen-value for all factors are acceptable b EFA for all
Following the elimination of items via Cronbach’s alpha analysis and exploratory factor analysis (EFA), the results indicated a KMO value of 0.631, which exceeds the acceptable threshold of 0.5 Furthermore, the Rotated Factor Matrix revealed that all items had loadings greater than 0.5, confirming their acceptance for further analysis.
The paper employed the SPSS software to divide behavioral factors into 5 components as is presented in Table 4.5
All loadings of items cluster were greater than 0.5 and met requirements to group them into 5 components The paper finished exploration factor analysis and used these components for regression analysis
Table 4.5: Summary of KMO and Bartlett’s Test, total variance explained and rotated component matrix (EFA time 2)
Initial Eigen values 1.658 (component 5) >1=> met requirement
Rotation sums of squared loading 59.924% >50% => met requirement
The study utilized SPSS software to identify five components, as illustrated in Table 4.7, with all item loadings exceeding 0.5, confirming their suitability for grouping Following the exploratory factor analysis, these components were employed in regression analysis Detailed results of the Cronbach’s alpha analysis at time 2 are provided in Appendix 2.
Table 4.6: Result of joint factor analysis for all scales
Testing the assumptions of regression
According to Leech, Barrett, and Morgan (2005), regression analysis serves as a statistical tool to examine relationships among variables In my research, I gathered data on various variables and utilized regression to assess the influence of independent variables on the dependent variable I chose the adjusted R², which is typically lower than the unadjusted R², to reflect the percentage of variance explained by the independent variables, noting that this adjustment is affected by both the effect size and sample size Additionally, I employed the F-test to evaluate the significance of the model, with hypotheses H₀ and H₁ outlined for analysis.
H 1 : At least one (at least one independent variable affect Y)
If p-value < 0.05, H 0 would be rejected while H 1 is accepted and the conclusion is that the research model is suitable
The standardized beta coefficient is interpreted similarly to the correlation coefficient, indicating the strength and direction of the relationship between independent and dependent variables A t-value less than 0.05 signifies that the beta coefficient is statistically significant, allowing us to assess whether the impact of independent variables on the dependent variable is positive or negative Additionally, Collinearity Statistics, specifically VIF and Tolerance, are important for identifying multicollinearity A low Tolerance Value, defined as Tolerance = 1/VIF, suggests the presence of multicollinearity among the variables.
1 – R 2 ), then there is likely a problem with multi collinearity.
Results of hypothesis testing
Table 4.7 Summary results of the hypotheses
H1 Risk tolerance has a negatively effcect on herding bias 0.109 0.086 > 05 Un-
Overconfidence and Gambler’s fallacy have a negative effect on herding bias
H3 Self –monitoring has a negative effect on herding bias 0.092 0.153 >.05 Un-
H5 Illusion of control has a positive effect on herding bias 0.358 0.000