Nghiên cứu về cảm tính nhà đầu tư trên thị trường chứng khoán việt nam TT TIENG ANH

MINISTRY OF EDUCATION AND TRAINING
UNIVERSITY OF ECONOMICS HO CHI MINH CITY

DOAN THI CAM VAN

RESEARCH ON INVESTOR SENTIMENT ON
VIETNAMESE STOCK EXCHANGE

Major: Finance - Banking
Number: 9340201

DISSERTATION SUMMARY

Academic Advisors
1. Assoc. Dr. Truong Thi Hong
2. Dr. Tran Thi Mong Tuyet

Ho Chi Minh City - 2022


This dissertation was completed at: University of Economics Ho Chi
Minh City
Academic Advisors:
1. Assoc. Pro. Dr Truong Thi Hong
2. Dr. Tran Thi Mong Tuyet
Independent Academic 1:………………………………………
Independent Academic 2:……………………………………….
Independent Academic 3: ……………………………………….
This dissertation will be defended at University level dissertation
committee: ……………………………………………
At…………………… day ………. month ………. year………

The dissertation can be found at the library: ……………………


LISTS OF PUBLISHED RESEARCH RESULTS
1. Assoc. Prof. Dr. Truong Thi Hong and Dr. Tran Thi Mong
Tuyet, M.Fin Doan Thi Cam Van, Dr. Le Long Hau “The effects of
the change in investor sentiment on stock returns on Hochiminh Stock
Exchange”, 3.2018, Vietnam Trade and Industry Review, vol 3/2018;
ISN: 0866-7756.
2. M.Fin Doan Thi Cam Van, Prof. Dr. Truong Thi Hòng, Dr.
Tran Thi Mong Tuyet, “The effects of investor sentiment on stock
returns”, Asian Journal of Economics and Banking, 10/2019, Vol 163;
ISSN: 1859-3682.
LISTS OF RESEARCH PROJECT
M.Fin Doan Thi Cam Van, Dr. Le Long Hau, M.Eco. Bui Le
Thai Hanh, 2018, “Assessing the impact of investor sentiment on stock
returns on Ho Chi Minh Stock Exchange”, Leader Research – M.Fin
Doan Thi Cam Van
CONFERENCE
Thi Hong, Truong and Thi Cam Van, Doan 2020, “Predictability
of investor sentiment” Conference 2020: Contemporary Issues in
Banking and Finance: Sustainability, Fintech and Uncertainties,
University of Economics Ho Chi Minh City.


ABSTRACT
Investor sentiment is a concept of behavior finance that always
exists in individual investors. Its approach helps to measure the level
of investor’s psychology. There are not many studies to explore the
effects of investor sentiment in the frontier stock markets although

these effects have been accepted in the developed and emerging stock
markets. Regarding Ho Chi Minh Stock Exchange (HOSE), a frontier
stock market, there has been still no research on investor sentiment
and analysis of its impact. Hence, this study will fill that research gap.
The results of this study show that the investor sentiment index of
HOSE always includes: number of stocks listed for the first time
(NFDL), turnover (TOR), the year-end ratio of the value weighted
average market-to-book ratios of payers and nonpayers (CMB), the
ratio of the number of advancing issues to declining issues
(ADC/DEC). Eliminating the phase of the market fluctuating
extraordinarily, the HSI’ composition has a new proxy that is the
average returns of stocks on the first day listing. The trading value of
foreign investors, the characteristic of frontier stock markets, is also a
component of the HSI. There is statistical evidence of the impact of
investor sentiment on the excess stock returns. There is also negative
correlation between the current investor sentiment state and the future
stock returns at the time t+3. HSI can predict the future returns on
stocks that are not difficult to arbitrage or to value stocks. The
distinctive characteristic of investor psychology on HOSE is aversion
to risk and prefer safe stocks. In conclusion, when the HOSE has come
into stable operation after its inception, the 6-factor model which is
adjusted for the impact of the macroeconomic factors is suitable to
measure the HOSE sentiment index.


1
CHAPTER 1. INTRODUCTION
1.1 Research motivation
1.1.1 Research situation of the world
Investor sentiment is a category of investor psychology. The

impact of investor sentiment and its predictability to stock returns have
been accepted on the developed and emerging stock markets of the
world. In the frontier stock markets that are strongly influenced by
investor psychology, there is not much research on the role of investor
sentiment.
1.1.2 Current situation of research and practice in Vietnam’ stock
market.
Ho Chi Minh Stock Exchange (HOSE), which is a frontier stock
market, has been demonstrated by many studies to be strongly
influenced by investor psychology. The current urgent requirement for
its development is to have solutions and technical tools to support the
market management and investment activities in this market. In
addition, there hasn’t been any research to measure psychological
level as well as to analyze the impact of investor sentiment on the
domestic stock market. Because of this current research gap, the study
on sentiment of investors on Vietnam’ stock market has been
conducted.
1.2 Research questions
1. The investor sentiment of HOSE market can be shown by which
trading statistic indicators? Is foreign trading value in the HOSE
market a proxy of the HOSE’ investor sentiment index?
2. How is the oscillation of the HOSE’ investor sentiment index?


2
3. Does the degree of the change in investor sentiment affect the
excess returns of stocks on the HOSE market? If so, in what direction
is that influence?
4. Does the change in the state of investor sentiment in HOSE
market (positive or negative) correlate with the future stock returns?

Can HOSE’s Investor Sentiment Index predict future stock returns?
1.3 Research objectives
1.3.1 Key research objective: This study measures the level of
investor sentiment and analyzes its effects on the Vietnam’ stock
market – the case of the HOSE market.
1.3.2 Detailed research objectives
1. Determining proxies constituting the investor sentiment index of
the HOSE market.
2. Measuring the level of the HOSE investor sentiment index.
3. Analyzing the level of investor sentiment change having impact
on excess stock returns.
4. Analyzing the influence of the investor sentiment state on future
stock returns and analyzing the ability to predict future stock returns
of investor sentiment index on HOSE market.
1.4 Scope and limitation of research
1.4.1 Research scope
The sentiment of investors who are trading on the HOSE is the
research scope of this study. They are all individual and institutional
investors. Another relevant research object is the stocks which are
listed in the HOSE.
1.4.2 Research limitation
1.4.2.1 The limitation of research space


3
HOSE is chosen because of its trading in stocks, the listed
corporations represent the national economics, and it is the biggest
stock market of Vietnam. HOSE is a tightly regulated and transparent
market, especially for information disclosure, so it is a good support
for measuring and analyzing the role of investor sentiment towards

Vietnamese stock market.
1.4.2.2 The limitation of research time
The HOSE’s size in the period of 2000 – 2005 was much smaller
than its size in the following years. There are many changes on the
market’s trading regulations because it is in the process of market
completion. So that, the information disclosure and collection of
market transaction data was not consistent in the regulation, affecting
the reliability of data. Based on the above reasons, the research period
is limited from 2006 to 2016, in which 2006 year is the base year, so
HOSE’s investor sentiment index and its effects are studied from 2007
to 2016.
1.5 Research methodology
HOSE’s investor sentiment is measured by the indirect method
offered by Baker and Wurgler (2006). Analyzing the effects of the
investor sentiment index is done by the descriptive statistical method
and regression analysis. Statistical tests were performed to detect and
handle hypothetical violations of regression for time series data such
as non-stationary data series, multicollinearity, heteroskedasticity,
self-correlation in regression analysis.
1.6 Scientific and empirical contributions
1.6.1 Scientific and academic contributions


4
The research results are the highly scientific reference source for
investor sentiment on the frontier stock markets in the world and in
Vietnam.
1.6.2 Empirical contributions
The study provides the investor sentiment index to measure
HOSE’s investor psychology, and to analyze the influences of this

index in the national stock market.
1.7 Research structure
Chapter 1: Introduction; Chapter 2: Theoretical basis and
literature review; Chapter 3: Data and research methodology; Chapter
4: The research results; Chapter 5: Summary, discussion and
implications
CHAPTER 2. THEORETICAL FRAMEWORK
AND RESEARCH DOCUMENTATION
2.1. Theoretical Framework
2.1.1. Traditional Financial Theory Framework
An overview of two important theories of the traditional financial
theory framework underpins this study.
2.1.1.1 Financial Asset Pricing Model of Fama and French Expansion Model
Experimental model of three factors affecting stock returns by
Fama and French (1993) and Fama and French model expanded by
Carhart (1997) are presented. On the basis of this model, investor
sentiment is expected to be a new factor added to the above model.
2.1.1.2 Efficient Market Hypothesis – EMH


5
That clearly understanding EMH theory in order to see its limitations
is a premise for developing subsequent theories.
2.1.1.3 Is the market really efficient? - The limitation of the EMH
theory
Evidence shows that a market is not always effective. The weakness
of traditional finance and EMH theory are the opportunity to develop a
theoretical framework of behavioral finance.
2.1.2 Theoretical framework of behavioral finance
There are two ways to explain the limitation of the traditional financial

viewpoint including the limit of arbitrage trading and the psychology of
an investor. This writing focuses on researching into investors’
psychology.
2.1.3 The sentiment of an investor is a psychological category
Conducting a research on psychology by detecting the bias in
investors’ perception cannot measure the level of psychology. There are
many types of cognitive bias that have been found and optimism is one
of them (Barberis and Thaler, 2003; Masomi and Ghayekhloo, 2010;
Chandra and Kumar, 2012). Optimism is an expression of emotions.
Emotion is a form of psychology that exists within each individual
investor (Baker and Wurgler, 2007). Thus, investor sentiment belongs
to the category of investor psychology. The advantage of doing research
on investor sentiment is to help measure the psychological level of
investors in the market.
2.1.4 General information about investor sentiment on the stock
market
2.1.4.1 General Information about investor sentiment


6
Synthesizing the concept of investor sentiment presented in
previous studies aims to bring out the concept of investor sentiment
within the scope of this study: “The sentiment of investors is the
perception, the expectation of investors on the prospect of the whole
market which is shown by the two states consisting of optimism and
pessimism”.
2.1.4.2 Measurement method
Currently, researchers, who research on sentiment, suggest two
main methods to measure investor sentiment.
Direct method: Conducting direct interviews with investors, who

participate in the market, in order to know their forecasts or feelings
about the current and future economic conditions or stock market
(Beer and Zouaoui, 2013). The advantage of this approach is that it is
not influenced by any theory. However, the representativeness of
investor sentiment across the market is the most controversial issue of
this method because it depends on the number of investors
participating in the interview. The control of respondents' emotional
concealment or their impressive tendency affecting interview quality
is also a notable issue for the direct approach.
Indirect method: Statistical parameters of the economy, the
market and the financial sector are used in order to capture investors'
thoughts and feelings about the current and future conditions of the
economy or stock market (Beer and Zouaoui, 2013). This method is
widely used (Baker and Wurgler, 2006, 2007; Kousenidis et al., 2011;
Changsheng and Yongfeng, 2012; Stambaugh et al., 2012; Boubaker
and Talbi, 2014, Deng et al., 2014, ... ) because of its objectivity, its
ease of implementation and its continuity of the measured investor


7
sentiment indicator. As a result, it is possible to grasp the level and the
change of investor sentiment. However, this set of indicators can be
theoretically debated because the suggestion of indicators reflect
sentimentality. These indicators may be influenced by other factors
(macro) from the economy.
The combination of these two methods will help limit each other's
weaknesses, which is a new proposed approach.
2.1.4.3 Factors expressing investor sentiment
Previous studies have mentioned several factors that can show the
emotions of investors. The writer has reviewed how the previous

authors have explained about the ability to display the investor
sentiment of a number of factors which are commonly used to express
the feelings of investors on the stock market.
2.1.5. Theory of the impact of investor sentiment on the stock
market
2.1.5.1. Hard-to-value stocks, hard-to-arbitrage stocks
Due to their own characteristics, it is hard for some types of stocks
to apply the pricing models. As a result, investors will find it difficult
to make investment decisions related to these stocks because of the
lack of valuation basis in terms of theoretical model. Therefore, these
kinds of stocks become quite sensitive to investor sentiment
(Glushkov, 2006). They are hard-to-value stocks. Investors are less
interested in them, and they are also considered as hard-to-arbitrage
stocks (Baker and Wurgler, 2007; De Long et al., 1990b).
2.1.5.2. Theory of the impact of investor sentiment on the stock
market


8
Regarding the theory of the impact of investor sentiment on the
stock market, the variation in price (or stock returns) of hard-toarbitrage stocks, hard-to-value stocks is in the same direction (in the
opposite direction) with the fluctuation of investor sentiment, and vice
versa for the group of safe stocks. In the meantime, the impact level
of investor sentiment is not the same for different types of stocks.
Hard-to-value stocks and hard-to-arbitrage stocks are more sensitive
to investor sentiment than safe stocks.
2.2 Summary of some studies on the influence of investor
sentiment on the stock market
2.2.1 Practical methods to measure sentiment of investors
The investor sentiment that is measured by a proxy is a single

sentiment index (Beer and Zouaoui, 2013; Sun et al., 2016; Li and Li,
2021, Phuong, 2020; Bui Kim Phuong, 2019). The investor sentiment
index may be also shown by several statistic indicators simultaneously in
which is called a composite index (Baker and Wurgler, 2006, 2007, 2012;
Berger and Turtler, 2012; Changsheng and Yongfeng, 2012; Aissia,
2016; Ding et al., 2018; Seok et al., 2019…). The components of the
investor sentiment index can be measured by direct or indirect
methods. To assess the influence of investor sentiment on the stock
market under different time condition, daily, weekly or monthly
investor sentiment indices have been established. Overall, it is
necessary to choose the appropriate methods to measure investor
sentiment for each stock market.
2.2.2 The influence of the investor sentiment on the stock market
2.2.2.1 The change of investor sentiment is a factor affecting excess
stock returns


9
There are several studies which found a positive correlation
between the change in the emotional level of investors and the excess
returns of stocks (Lee, 1991; Glushkov, 2006; Changsheng and
Yongfeng, 2012; Berger and Turtle, 2012). It means that if the
changing level of sentiment in a market increases, it will lead to the
increase of excess returns. Conversely, if the change is less, it will
result in the decrease of excess returns.
2.2.2 Investor sentiment affects future stock returns
The influence of sentiment on future stock returns is not completely
the same between the stocks having different attributes, or having the
same ones but different levels in characterization. There is an inverse
correlation between investor sentiment and future stock returns (Baker

and Wurgler, 2006, 2007). Many other authors have also confirmed
this (Kousenidis et al., 2011; Hribar and McInnis, 2012; Beer and
Zouaoui 2013; Corredor et al., 2014; Beer and Zouaoui; 2013).
2.2.3 Predictability of investor sentiment index on future stock
returns
When the investor sentiment is at a high level, this will lead to the
fact that it is hard to value the future returns of hard-to-value stocks,
hard-to-arbitrage stocks and the price difference will be lower than that
of the remaining future returns of hard-to-value stocks, hard-toarbitrage stocks. On the contrary, since the sentiment is low, it will be
difficult to value the future returns of hard-to-value stocks, hard-toarbitrage stocks, challenging to run business and the difference of the
price will be bigger than that of the other future stock return groups
(Baker and Wurgler, 2006). This conclusion is not only made in the US
stock market but also in other markets including Europe, India, China,


10
Hong Kong, Thailand, ... (Debata et al., 2017; Fintner et al., 2010 ; Lux,
2011; Kousenidis, et al., 2011; Sayim, 2012; Corredor et al., 2013;
Changsheng and Yongfeng, 2012; Jiang and Li, 2013; Majumder, 2014;
Kumari and Mahakud, 2015).
2.3. The gaps and the directions of filling the research gaps
1. Narrowing the research gap on investor sentiment on the frontier
stock market.
2. Measuring investor sentiment with a composite index rather
than a single index or gauge sentiment by investor behaviors.
3. Adding a new factor of the investor sentiment index which has
the characteristics of the frontier stock market - Trading value of
foreign investors on the stock market.
2.4 Hypotheses need to be examined in the research
Hypothesis 1: the proxies such as NFDL, RIPO, TOR, CMB,

ADV/DEC and trading value of foreign investors are components that
represent investor sentiment in the HOSE market.
Hypothesis 2: The change of investor sentiment is positively
correlated with excess stock returns.
Hypothesis 3: There is an opposite variation between investor
sentiment and future stock returns.
Hypothesis 4: Investor sentiment possibly has reversed forecast on
the future hard-to-value stocks, hard-to-arbitrage stocks returns.
CHAPTER 3. RESEARCH METHODS AND RESEARCH DATA
3.1 Research Process
Firstly, based on the technical background of principal component
(PCA) analysis, the writer determined the measurement model and


11
measured the level of investor sentiment of HOSE by an indirect
method. Next, the writer analyzed the sentiment changes affecting the
stock returns. Then, the writer continued to examine the variation of
the future stock returns within each current state of the investor
sentiment (optimistic or pessimistic). Finally, the writer analyzed the
influence of investor sentiment on the difference of the future stock
returns and looked for the predictability of investor sentiment index of
HOSE (HSI).
3.2 Theoretical research model
3.2.1 The model of measuring the emotional level of investors
𝐹𝑡 = ∑𝑚
𝑖=1 𝜇𝑖,𝑡 ∗ 𝑠𝑡𝑑(𝑋𝑖,𝑡 ) (3.1)
HSI is measured by an indirect method thanks to: (1) the
advantages of the indirect method; (2) the fact of lacking data to apply
other methods. Principal component (PCA) analysis has to be done

twice in order to filter and choose factors which are closely related to
each other and have sufficient statistical basis to show that they are
capable of expressing investor sentiment. This results in the
determination of the measurement model of investor sentiment. HSI
scores are determined by the following model:
𝐹𝑡 = ∑𝑚
𝑖=1 𝜇𝑖,𝑡 ∗ 𝑠𝑡𝑑(𝑋𝑖,𝑡 ) (3.1)
- Ft is the score of the investor sentiment index (new factor) at
the time t.
- m is the remaining sub-variable after the first PCA analysis (m
≤ n / 2); n is the number of measurement variables proposed to be
included in the model; i,t is the component point coefficient of the ith
measurement variable at the time t, this is the value of the eignevetor
of each observed variable.


12
- Std (Xi, t) is the standard deviation of the measurement
variable Xi, t.
3.2.2 The model of analyzing the emotional effects of investors on
the stock market
This analysis is based on the theoretical multi-factor model and the
experimental model of Fama and French (1993) and is extended by
Carhart (1997). In which, the factor of investor sentiment is expected
to be a new factor in the model.
3.3. Experimental research model and research method
3.3.1 Measuring investor sentiment index of HOSE
This is the first study on the issue of investor sentiment on
Vietnamese stock market. Therefore, the proposed factors are selected
based on the research results of Baker and Wugler (2006) and some

previously published studies on other stock markets. In addition, the
writer also added a new factor variable - the transaction value of
foreign investors - which brings specific characteristics of the frontier
stock market of Vietnam into the measurement model of the investor
sentiment.
3.3.1.1 Factors included in the HSI measurement model
They are number of stocks listed for the first time (NFDL); Stock
returns on the first day of listing (RFDL); turnover (TOR); The yearend ratio of the value weighted average market-to-book ratios of
payers and nonpayers (CMB); The ratio of the number of advancing
issues to declining issies (ADV/DEC); and trading value of foreign
investors (FOT). In order to meet the research objectives, the new
issuance timeline is changed to first listing to suit the characteristics


13
of HOSE market, which is the different issuance time compared to the
listing time of stock.
Regarding trading value of foreign investors, several studies
conclude that trading value of foreign investors affects domestic stock
market (Kang et al., 2016; Lim et al., 2016; Dhingra et al., 2016; French,
2017). In addition, the transaction of foreign investors on HOSE has a
strong influence on the price of stocks in the market (Vo, 2017). There
is an evidence of a possitive unusual net buying day of foreign investors;
however, there is no clear evidence of their negative unusual net sale
day (Dang Buu Kiem, 2018). These results has raised the suspicion that
the transactions of foreign investors will be paid more attention by
domestic investors because foreign investors are considered to be
experienced investors; therefore, domestic investors tend to follow the
foreign investors. As a result, when the increase in foreign investors'
trading value increases, it is expected that investor sentiment in the

market would become more optimistic and vice versa.
Market transaction statistics may reflect investor sentiment at the
same period or one period later (Baker and Wurgler, 2006). Therefore,
with the original 5 factors (or 6 factors), the proposed model will have 10
factors (or 12 factors) selected to establish the HSI measurement model.
3.3.1.2 The experimental model to measure investor sentiment
index of HOSE
The empirical research model of HSI measurement has the
following form:
𝑧
𝐹𝑡 = ∑𝑧′
𝑖=1 𝜇𝑖,𝑡 ∗ 𝑠𝑡𝑑 (𝑋𝑖,𝑡 ) + ∑𝑖=1 𝜇𝑖,𝑡−1 ∗ 𝑠𝑡𝑑 (𝑋𝑖,𝑡−1 )

(3.3)

z and z 'are root factors (t) or (t-1) remaining after PCA analysis.
Xi is the proposed factors for the mentioned model in section


14
3.3.1.3 Steps to measure investor sentiment index of HOSE
In order to start the 1st PCA analysis, the model includes the
proposed factors to determine the intermediate HSI (HSI*). The model
is suitable when the KMO coefficient of each factor is greater than>
0.5 and the whole model is ≥ 0.6. Next, the writer determined the
correlation coefficient between the components of HSI* with HSI* to
serve as a basis for refining the factor (Xi, t or Xi, t-1) provided that
the retained factor must have higher correlation coefficient. Then, the
writer did the 2nd PCA analysis towards the retained factors to
determine the official HIS. This is the first component of PCA result.

However, the final model which is selected has to ensure the same
conditions of KMO test coefficient as the previous analysis. In
addition, the correlation coefficient between HSI and HSI * must be
greater than 0.9 in order to ensure that the elimination of the variable
does not cause data loss.
Next, the writer estimated the HSI score. If the HSIt score is higher
than (>) the mean of the HSI during the study, it means that investor
sentiment is at high level and is considered as in an optimistic state. If
HSIt is less than (<) the mean, it means that investor sentiment is at
low level is considered as in a pessimistic state.
3.3.2 Analyzing the effect of the change of investor sentiment on
excess stock returns
The proposed experimental model is as follows:
R p,t − R𝑓,t = 𝛼0 + 𝛾1 ∆HSIt−k + 𝛽1 RMKTt + aSMBt + bHMLt +
cUMDt + up,t

(3.5)

Investors need time to receive and process information. The time
to receive and process information will be different for each


15
individual, community,... However, this process cannot last too long
in the context of the current economy. This is the reason why the study
only investigates the effect of ∆HSIt = HSIt – HSIt-1 on the stock
returns that stopped at the maximum time t + 3 (Rt, Rt + 1, Rt + 2 and
Rt + 3 ).
3.3.3. Analyzing the correlation between investor's emotional state
and future stock returns

Firstly, the writer set the stock portfolios based on company
characteristics with 5 levels (level 1 is the lowest and level 5 is the
highest) at the beginning of each fiscal year. Next, the author
calculated the average of monthly stock returns of each stock
portfolios. Finally, the writer made the statistics of their stock returns
right at the time t + k (k = [1, 2, 3]) based on the status of investor
sentiment (optimistic/pessimistic) at the time t, and then made survey,
evaluated the variation between them.
The characteristics of the company including company size,
company value (BE/ME), listing time, volatility, company capital
structure, asset structure, business performance, and investment
efficiency are surveyed.
3.3.4. Experimental model analyzes the predictability of investor
sentiment
The analysis of the impact of HSIt on the difference in future stock
returns between stock portfolios is done by the model:
R X =high,p,t − R X= low,p,t = α1 + γ2 HSIt−k + 𝛽2 RMKTt + dSMBt +
eHMLt + gUMDt + εp,t

(3.17)


16
Combining the results of regression analysis and the variation
between future stock returns and the HSIt status to detect the
predictability of the HSI index.
3.4. RESEARCH DATA
3.4.1. The source of research data
The stock market transaction statistics and financial indicators of
each stock code are listed on HOSE, which were collected from

January 2006 to December 2016. This data is officially provided by
Tai Viet Joint Stock Company (Vietstock) with a guarantee of an error
that it does not exceed 2%.
3.4.2. An overview of research data and characteristics of
companies listed on HOSE
In order to have an overview of the data set, the writer made
statistics to describe the characteristics of research data as well as
evaluated the company characteristics of stocks. The evolution of
listed companies is the object of this study.
CHAPTER 4. RESEARCH RESULTS
4.1 Expression factors and measurement models of investor
sentiment on HOSE (2007 – 2016)
4.1.1 Investor sentiment measurement model on HOSE (2007 – 2016)
Both original factor model (HSI_0716) and controlled factor
model (HSI_0716┴) have the same four factors, include: number of
stock listed for first time (NFDLt), turnover (TORt), the year-end ratio
of the value weighted average market-to-book ratio of payers and
nonpayers (CMBt-1) and the numbers of advancing issues to declining
issues (ADV/DECt-1). The trading value of foreign investors (FOTt-1)


17
is a factor of both the HSI_F0716 model and the HSI_F0716┴. All of
these factors are positively correlated with the investor sentiment
index.
The essential difference of the above investor sentiment
measurement models compared to the Baker and Wurgler’ model is
the absence of RFDL factor in those models. Besides, the positive
correlation between CMB and HSI is also the other difference from
Baker and Wurgler’ model (2006, 2007). The difference in preference

between the investors in the HOSE and the investors in the developed
stock markets is considered to be the cause of the above contradictory
results.
Removing the impacts of the macro variations from HSI’s proxies
reduces the proxy’s devotions to the model and its interpretability.
Adding the FOTt-1 to the removed model did not cause changes in
HSI’s variability, but there is not enough evidence to conclude which
model is better.
4.1.2 Investor sentiment measurement model on HOSE (2010 – 2016)
The models (2010 – 2016) were formed to appraise whether the
period of abnormal fluctuations (2007 – 2009) could affect the models
to measure investor sentiment or not. Results show that the models
(2010 – 2016) still have some similarities with the models 2007 – 2016
such as NFDLt, CMBt-1, ADV/DECt-1 are still proxies of investor
sentiment index, the correlation between each of the above proxy and
investor sentiment index are still positive.
TORt-1 factor reflects investor sentiment at time t in the models
(2010 – 2016) is the change compared to the models (2007 – 2009) in
an appropriate direction, similar to the model of Baker and Wurgler


18
(2006). RFDLt became an indispensable proxy of the investor
sentiment measurement models in the period 2010 – 2016.
Cleaner models with proxies that eliminate the effects of macro
factors have higher coverage than the models with raw proxies.
Despite a change in investor sentiment scores, its variable trend is
unchanged. FOTt-1 is still a proxy of investor sentiment and the
HSI_F1016┴ has the highest significant level compared to other
models (KMO = 0.74).

Thus, the market period of abnormal fluctuation affects the HSI’
measurement models. Investor’ behaviour is changed to a more
appropriate direction, which approves that they have more experience
in investing activities on the stock market. Macro variations also affect
the HIS’ measurement models so it should be to eliminate the
influence of macro indicators to the proxies of investor sentiment
before measuring HSI index. Macro indicators influence HSI’
measurement models, so it should be to remove their effects on the
proxies of investor sentiment before measuring HSI index. The trading
value of foreign investors is a proxy of the HSI models in this period.
There are not enough bases to choose the most suitable model.
4.2. Variation of investor sentiment on HOSE market
4.2.1 From 2007 – 2016
Value 0 is the boundary between the optimistic and pessimistic
state of HSI. When HIS>0, the sentiment state is optimistic.
Conversely, this is pessimistic.


-4

-2

0

2

4

19

2006m1

2008m1

2010m1

2012m1
time

HSI_0716
HSI_F0716-

2014m1

2016m1

HSI_0716-

Resource: Results of PCA analysis
Picture 4.7 The HSI indexes (2007 – 2016)
Investor sentiment was in a high optimistic state almost during
2007 – 2010. When the market situation is more stable, the state of
investor sentiment is not as high as before. After year 2010, although
there were almost of the time when investor sentiment was low level
(pessimistic), there were still sometimes when investor psychology
became more positive, i.e. investor sentiment moved to high level
(optimistic).
4.2.2 From 2010 – 2016
Investor sentiment state was optimistic almost all year 2010.
HSI_1016┴ and HSI_F1016┴ showed signs of a reversal in investor

sentiment in the last few months of 2010. In the following years,
HSI_1016 was always low (pessimistic) but HSI_1016┴ and
HSI_F1016┴ sometimes turned optimistic, though not for too long.
The variation of investor sentiment indexes in the period 2010 – 2016
is equivalent to the variation of them in the last years of the period
2007 – 2016.


20
In summary, HOSE’ investors have gradually understood and
accumulated investment experience in the stock market. In the last
years of the research period, the market no longer appeared to have an
excessive psychological reaction of investors as before.
4.3 The change of the investor sentiment affect to excess stock returns
4.3.1 From 2007 – 2016
The effects of the conditional factors such as market risk (RMRF),
firm size (SMB), firm value (HML) and momentum stock (UMD) on
the excess stock returns in the regression analysis model are
statistically significant which is appropriately explained in the actual
market of HOSE. With the change of sentiment factors such as
∆HSI_0716t; ∆HSI_0716┴t và ∆HSI_F0716┴t:
1. There is evidence of the effects of ∆HSI on the excess stock
returns in all stock portfolios. (2) The effects on the excess stock
returns appear at two times and have a reversal of the effects at those
two times. It mean that, all three indexes, including ∆HSI_0716t;
∆HSI_0716┴t and ∆HSI_F0716┴t have positive effect (+) on the stock
returns at time t, but ∆HSI_0716┴t and ∆HSI_F0716┴t have negative
effect (-) at time t+2. (3) The change of investor sentiment index has
a more significant impact in the present than in the future.
4.3.2 From 2010 – 2016:

During the stable period of the market, the effects of the factors
such as RMRF, SMB, HML, UMD on the excess stock returns is more
and more obvious.
The

influence

time

of

∆HSI_1016t;

∆HSI_1016┴t and

∆HSI_F1016┴t is delayed with the lag is one and the sign of the impact
of them is lower than in the period 2007 – 2016. The correlation


21
between the change of investor sentiment and the excess stock returns
is positive (+) in the short term and negative (-) in the long term.
Removing the impact of macro indicators from the proxies of investor
sentiment index and adding the FOT factor to the investor sentiment
measurement model shows the effect of the ∆HSIt-k factor on the
excess stock returns is reduced.
In conclusion, ∆HSIt is a new factor that has a significant effect on
the excess returns. The effect on the excess returns of the investor
sentiment change is stronger than the RMRF. The market
characteristic is a reason for the term difference of the influence; the

impact of the investor sentiment change in the stable market period is
lower than in the unstable market period. Investor sentiment index
which is measured by the model that omits the impact of macro factors
correctly reflects the market fact that the excess returns are less
affected by investor sentiment when the stock market is stable.
4.4 The impact of investor sentiment state on the future stock returns
4.4.1 From 2007 – 2016
The Rt+k of portfolios with HSIt = optimistic is lower than the Rt+k
of portfolios with HSIt
optimistic and HSIt

= pessimistic.

= pessimistic,

In addition, when both HSIt =

HOSE’ investors still tend to prefer

securities with safety characteristics because Rt+k of these securities
are always higher than Rt+k of the securities with opposite
characteristics. Those safety stocks are large firm size stocks, high
firm value stocks, high volatility stocks, low external finance stocks,
high fixed assets stocks, profitable or high profitable stocks, dividendpaying or high dividend-paying stocks.
4.4.2 From 2010 – 2016