Empirical Research of January Effect in Vietnam Securities Market

Graduate School of Finance, Shu-Te University Empirical Research of January Effect in Vietnam Securities Market Student: Do Van Son ABSTRACT The January effect implies that securities

Shu - Te University

College of Management Graduate School of Finance

Empirical Research of January Effect in Vietnam

Securities Market

Student : Do Van Son

Advisor : Dr Juping Wu

Dec, 2014

Empirical Research of January Effect in Vietnam

In Partial Fulfillment of the Requirements

For the Degree of

Master of Finance

Dec, 2014

Graduate School of Finance, Shu-Te University Empirical Research of January Effect in Vietnam

Securities Market

Student: Do Van Son

ABSTRACT

The January effect implies that securities market experience in abnormal return in January In this thesis, I conduct the empirical research of the January effect on monthly returns of Vietnam securities market, specifically VN-Index The research is aimed to find out the existence of January effect on overall period, subperiods of VN-Index and on securities portfolio with different market capitalization By using the method of least square, the research result has shown that January effect exists overall period of VN-Index with higher effect during uptrend period but there has no effect during downtrend period The effect also does not exist on different market-cap portfolio during the downtrend period The observation of this effect can help investors to establish a profitable investment strategy and effective risk management

Keywords: January effect; Abnormal returns; Vietnam securities market;

Emerging Market

i

I would like to take this opportunity to recognize the contribution of the people who supported me in completing this thesis Without their valuable support and guidance, this thesis would not have been possible

Firstly, I would like to express my gratitude to professors of the University SHU-TE and the Foreign Trade University, who taught and helped me throughout the course Special thanks to my Academic Advisor Dr Juping Wu and Co-advisor Assoc.Prof.Dr Nguyen Thu Thuy, who provided very useful guidance and recommendation throughout the process of writing this thesis Their support is of vital importance to the successful completion of this thesis

Secondly, I would like to thank to my MBAF4 companions and friends for their sharing and helping me during the study at the MBA Program and for the

completion of this thesis

Finally, I would like to send my deepest gratitude and love to my wife and sons Without their encouragement and continued support, I would not have been able to complete the dissertation and the whole two years of studying the MBA program

Đỗ Văn Sơn – MBAF4, Shu-Te University, 2014

ii

TABLE OF CONTENT

ABSTRACT i

ACKNOWLEDGMENTS ii

LIST OF TABLES v

LIST OF FIGURES vi

ILLUSTRATION OF SYMBOLS vii

CHAPTER 1 - INTRODUCTION 1

1.1 Research background 1

1.2 Research questions and objectives 5

1.3 Scope of the research 5

1.4 Motivation of the research 5

1.5 Research procedure 7

CHAPTER 2 - THEORETICAL BACKGROUND 8

2.1 Overview of the January Effect 8

2.1.1 January Effect in securities markets 8

2.1.1.1 January Effect in Emerging Securities Market 8

2.1.1.2 January Effect in Developed Securities Market 9

2.1.1.3 January Effect is not unique to small firm but it is more statistically significant with small firms 10

2.1.2 Reasons for the January Effect 11

2.2 The concept and formula of Securities Index 12

2.2.1 The concept of securities index 12

2.2.2 The Formula for VN – INDEX 13

CHAPTER 3- RESEARCH METHODOLOGY 15

3.1 Research models for January effect 15

3.1.1 Non-linear pattern by a Taylor expansion 15

3.1.2 OLS methodology by Cristina Balint and Oana Gica (2012) 16

3.1.3 Time series Garch framework 16

3.2 Estimation of research model, description and variable measurement 16

iii

3.2.1 Return calculation formula 16

3.2.2 Application of regression model to test the January effect on VN-Index 17

3.2.3 Application of regression model to test the January effect on VN-Index’s sub-periods 18

3.2.4 Application of regression model to test the January effect on company’s market capitalization 19

3.3 Research data 20

3.3.1 Securities market data 20

3.3.2 Company data 20

CHAPTER 4– DATA ANALYSIS AND RESEARCH RESULTS 23

4.1 The January effect on VN-Index in the period 2000 – 2013 23

4.2 The January effect on VN-Index in the period 2000 – 2007 25

4.3 The January effect on VN-Index in the period 2008 – 2013 27

4.4 The January effect on Large-cap Portfolio during period 2008 – 2013 28 4.5 The January effect on Mid-cap Portfolio during period 2008 – 2013 30

4.6 The January effect on Small-cap Portfolio during period 2008 – 2013 33 CHAPTER 5 – CONCLUSION 35

REFERENCES 38

ANNEX 41

iv

LIST OF TABLES

Table 1: Average returns by months of VNI during 2000 - 2013 13

Table 2: Market capital classification 21

Table 3 - Large-cap securities portfolio 21

Table 4 - Mid-cap securities portfolio 22

Table 5 - Small-cap securities portfolio 22

Table 6 - Tabulation of VN-Index Return in period 2000 - 2013 24

Table 7 - Regression model for VN-Index returns in period 2000-2013 24

Table 8 - Regression model for VN-Index returns in period 2000-2007 25

Table 9 – Lei Gao model for VN-Index returns in period 2000-2007 26

Table 10 - Regression model for VN-Index returns in period 2008-2013 27

Table 11 – Lei Gao model for VN-Index returns in period 2000-2007 28

Table 12 - Regression model for large-cap portfolio 29

Table 13 – Lei Gao model for Large-Cap Portfolio 30

Table 14 - Regression model for Mid-cap portfolio 31

Table 15 – Lei Gao model for Mid-Cap Portfolio 32

Table 16 - Regression model for Small-cap portfolio 33

Table 17 – Lei Gao model for Small-Cap Portfolio 34

v

LIST OF FIGURES

Figure 1: Return of Securities from 2000 to 2007 1

Figure 2: Return of Securities from 2008 to 2013 2

Figure 3: VN-Index performance in period 2000-2013 14

Figure 4: VN-Index histogram and descriptive statistics 23

vi

𝑤𝑤𝑖𝑖 : Weighted number of stock

𝑅𝑅𝑖𝑖 : Return rate of stock i

m : Total number of stocks in the portfolio

s

P : Stock price at time (t – 1)

vii

CHAPTER 1 – INTRODUCTION

1.1 Research background

Vn-Index officially operated since 27th July 2000 with the initial index of

100 points and reach to 571 points in 25 June 2001 By that time, there are a few listed companies in the market By the end of 2001, VNI went down to 235,4 points, and then 130 points in October 2003, the lowest point in VNI’s performance history

Hereinafter are the figures showing monthly returns of securities before the economic recession (2000-2007) and during economic recession (2008-2013)

in Ho Chi Minh stock exchange from 2000 to 2013

Figure 1: Return of Securities from 2000 to 2007

(Source: HOSE, own calculation)

The period 2006-2007 is considered as the “golden time” for VNI By that time, VNI reached to the highest point (around 1.138 points in February 2007) in history of this market It is due to Vietnam’s joining to WTO, preferred

tax for listed companies, many good companies were listed…

Figure 2: Return of Securities from 2008 to 2013

(Source: HOSE, own calculation)

The period from 2008 to 2013 is considered as the down-trend period, especially the period 2008-2009 is known as a “nightmare” to almost investors For example, VNI decreased in 7 consecutive trading days, wiping out 88,58 points, equivalent to 21% The main reasons for this tragedy are financial crisis

in the US, debt crisis in EU countries and uncertain economy in Vietnam

From the performance of Vietnam’s market, it can be seen that this market is not efficient since investors behave on crowd-effect basis For example,

on 25th November 2009, Vietnam’s State Bank has decision on adjustment of foreign exchange rates and increase basic interest rate to 8% after 11 months maintaining at 7%, so investors sold very aggressively and VNI decreased by 8,18% only in two days Therefore, the understanding behavior of investors in the market is very important for winning this market

Inspite of its high fluctuation, Vietnam Securities Market is still an important investment channel for investors, from professional investment institutions to individual investors Therefore, it is very significant for

researchers in terms of the finding of its laws and signals to help market participants have the base for making investment decision, state management agencies have the base for policy issuances and market development orientations One of the issues that many researchers around the world, except Vietnam’s researchers, pay much attention is abnormal return in securities market

Returns generated by a given security or a portfolio over a period of time, different from the expected rate of return or not conform to the law of Efficient Market Hypothesis is known as abnormal returns The Efficient Market Hypothesis confirms that abnormal return does not exist in securities markets because all information is reflected in stock prices However, many investors are trying to seek for abnormal return by learning on movement pattern of stock prices in the past The fact has shown that there exist such opportunities to earn abnormal returns thanks to the taking advantage of some laws or seasonal effect

in securities market Seasonal effects that many people are interested in are daily effect of a week, January effect, May effect, October effect One of the effects that many researchers are especially interested is January effect

The January effect describes the abnormal return of stock prices in several weeks at the end of December and at the beginning of January in next year For example, in January 2013, almost 20 largest markets’ index in the world increases (except Brazil and Korea) Among those markets, Japan, China, UK, Switzerland and Italy’s indices has higher increase than that of the US Specifically, in January 2013, Nikkei index of Japan went up by 7.2%, Composite Shanghai (China) up by 6.5%, S&P500 (US) up by 5.5% January effect has been significant since 1950 in the US, because in past 63 years, only in

1987 this effect is not shown in the market

In previous special reports, Rozeff and Kinney (1976) initially find that the existence of seasonal laws of stocks and most of those laws prevail in January This finding has created a new research path for finding the existence of the January effect, its reason and its conflicts with Efficient Market Hypothesis

3

(EMH) The reasons for hypothesis of abnormal return is rather abundant, for example, Tax avoidance hypothesis, Windows dressing hypothesis and Liquidity hypothesis

Those above reasons have partly explained the abnormality of stock returns but there is no true reason for January effect’s conflict with Efficient Market Hypothesis Therefore, the finding of a main reason for the explanation purpose of those effects continues Especially, investors in the US market have more than five decades taking advantage of this effect to earn abnormal returns

In Vietnam securities market, there are some common reasons as stated in hypothesis above Regarding the Liquidity hypothesis, Vietnam’s companies usually pay salaries and bonuses at the end of the year as a very important manner to encourage employees to contribute intensively in a new year Regarding the Windows dressing hypothesis, there are an abundant number of public funds and private funds in Vietnam market that usually apply techniques

to “polish” their financial reports by selling unrealized-loss stocks at the end of the year and buy it back at the beginning of next year In Vietnam, The Vietnam state Bank and commercial Banks usually expand credit intensively at beginning months of the year, especially individual loans, so many investor expect such loans to be invested into securities market Besides, positive information to support the market usually announced at the beginning of a year, thereby, having

a positive impact on securities index The abnormality of stocks’ performance in January has caused a conflict with Efficient Market Hypothesis For such reasons, the surveying and researching on January effect in Vietnam securities market is quite feasible and bring many benefits to market participants The control of those abnormalities are essential, especially it should be controlled by state

management agencies Therefore, the thesis “Empirical Research of January

Effect in Vietnam Securities Market” is implemented to find out and justify both

theoretical and practical in Vietnam securities market

4

1.2 Research questions and objectives

This research aims to solve the following issues:

1 Does January Effect exist in Vietnam Securities Market?

2 What is the impact of January Effect on sub-periods of Vietnam Securities Market (before economic recession and during economic recession periods)?

3 What is the impact of January Effect on small-cap, mid-cap and large-cap stocks in Vietnam securities market?

1.3 Scope of the research

This study examines Vietnam Securities Market and companies during the period from September 2000 to December 2013 The research data is monthly returns of VN-Index and company stocks from September 2000 to December

2013 Company data shall be collected accordingly with starting time of stocks listed in Hochiminh Stock Exchange (HOSE)

1.4 Motivation of the research

With the application of financial models and Behavioral Finance Theory used worldwide, the January effect has been tested in developed and emerging markets Vietnam securities market just has a development history of over 10 years; therefore, state management agencies, individual investors and institutional investors do not have sufficient knowledge and experiences At this time, there is no thorough research on this seasonal effect in Vietnam securities market Therefore, the research on the existence and its reasons in Vietnam securities market has a very important significance in terms of both theorical and practical to whom its may be concerned The research result is expected to contribute and provide important findings on January effect and meet the following objectives

From academic perspective, this research shall provide additional knowledge on January effect in Vietnam securities market and its evidence for the existence or non-existence of January effect in Vietnam, especially the

5

increasing of small-cap stock prices in January from year to year This research shall provide empirical evidence for the comparison of Vietnam’s January effects with other emerging and developed markets Moreover, its research results may also be utilized as a supplemental evidence for present researches on January effects and provide useful information and findings for other future researches and it is provide the basic research about the continue seasonal effect research on Vietnam stock market, as well

The seasonal pattern of market returns violates the assumption of weak market efficiency in that by observing the past development of returns, market participants can make extraordinary profit Therefore, the perception of the January effect existence and its law shall help investors to make wise investment decisions, have a good timing of investment and have right strategies to generate abnormal returns from this seasonal effect, thereby reducing risk of losing money

on their investment portfolios

It is important to know the impact of January effect on different periods of the market The impact of January effect in high-growing economy is usually higher than that in low-growing economy; therefore, investors should have good strategy on such different stages or periods

In addition, the impact of January effect may be different due to the market capitalization of company Statistical data has shown that January effect usually has higher impact on small-cap companies; hence, investors should be aware of this difference

The January effect is one of seasonal effects reflecting the inefficiency of securities market To have an efficient market, it is important that state management agencies should know about the reason for the existence of January effect The lack of transparence, price control, insider trading… still exists in the market This understanding shall facilitate them to issue sound policies for market development and efficiency improvement of Vietnam securities market

6

1.5 Research procedure

As we know that this effect is significant in global securities market Therefore, many foreign researchers have found out models to prove the existence of this effect This thesis’s main goal is also to find the evidence for the January effect in Vietnam

Chapter 1: Introduction

Chapter 2: Theoretical background

To study the January effect in the world and its reasons for January effect The research methodology and data collect from Vietnam Securities market is introuduced in

Chapter 3: Research Methodology

Providing a suitable model from tested models widely used by researchers from China, Romanie, USA, etc.shall be selected for its application to Vietnam’s market data Specifically, the January effect on different stages of the securities’ index and on different market cap portfolio shall be tested with the quantitave model The finding of this research shall be very useful and significant for both business and academic fields in Vietnam

Chapter 4: Data analysis and research results

To provide detail results of empirical test for the January effect on different periods and market-cap potfolios

Chapter 5: Conclusions

To provide for lesson learnt, its significance and future research issues

In next chapter, the literature review shall be done in order to learn about the study results of worldwide researchers, through which we can know how the January effect on different markets and the reasons for its existence This analysis shall be the basis for the research of the January effect in Vietnam’s securities market

7

CHAPTER 2 - THEORETICAL BACKGROUND

2.1 Overview of the January Effect

2.1.1 January Effect in securities markets

The January effect, also known as the turn-of-year effect, is a calendar effect wherein securities increase in value more rapidly than in other months Although in Vietnam there is no thorough research on the January Effect, there have many studies concerning the January Effect in the world

2.1.1.1 January Effect in Emerging Securities Market

Wong, Agarwal and Wong (2006) state that in the pre-crisis period 1997), the January effect exists in Singapore market However, analysis in the post crisis period shows that such anomality has significantly declined or disappreared

(1993-Chung-Wen (2005) states that Taiwan index and Hongkong index show high probability of the existence of January effect

Wong, Ho and Dollery (2007) test monthly effect in Kuala Lumper Composite Index for the period from January 1994 to December 2006 For the whole period, the monthly effect does not exist However, when the three sub-periods are examined, the regression results confirmed a monthly effect in stock returns It is found that the results are positive and significant for January and February during the post-crisis and pre-crisis period respectively The results also indicate the existence of negative March and September effects for the post-crisis period

Deyshappriya (2014) examines the Stock Market Anomalies in Colombo Stock Exchange, Sri Lanka during the period of 2004 to 2013 The existences of Monthly effect has been tested using monthly data The Ordinary Least Square (OLS) method were employed to capture the effect The sample period was divided into two periods as War Period and Post War Period in order to take the

8

impact of the War into account The result has showed that statistically significant positive January and negative December effects were observed for the War period and entire sample period However, the irregular monthly effect has been found for post War period with significant positive April, September and December effects but not January

Nageswari, Selvam, Vanitha and Babu (2013) use the logarithmic data for S&P CNX Nifty and S&P CNX 500 sample indices and applied the Dummy Variable Regression Model from 1stApril 2002 to 31st March 2011 The result indicates the highest mean return was earned in December and the lowest/ negative mean return earned in January Month for S&P CNX Nifty index The S&P CNX 500 Index recorded the Highest Mean Return in the Month of March and the Highest Negative Mean Returns in the Month of January It is also found that there was significant difference in the mean returns among the different months of the year The analytical results of seasonality indicate the absence of January Anomaly during the study period

Ahsan and Sarkar (2012) study the January Effect in Bangladesh stock market from 1987 to 2012 but the results did not show the January Effect In fact during the sample period, mean stock return from December to April was negative

2.1.1.2 January Effect in Developed Securities Market

Marret and Worthington (2011) do the empirical research in the Australian market to find out the monthly effect on Austrialian Stock Exchange Index’s return and industry returns At the market level, evidence is found of significantly higher returns in April, July and December (up to nearly three times higher than average returns across all months) It is interesting that the impact for small cap firms is even more significant with January, August and December returns being 5.3, 3.9 and 4.9 times higher than mean return throughout the year

At the sub-market level, monthly effects are found in the diversified financial, energy, retail, telecommunications and transport industries, but not in the

9

banking, healthcare, insurance, materials and media industries Of these, the most substantial month-of- the-year effects at the industry level relative to mean industry returns are January returns in the retail industry which are more than twice as low as returns in other months of the year and in the telecommunications industry where they are more than thirty-three times higher

Tylor (2007) tests the January effect across 16 countries (Australia, Austria, Belgium, Canada, Denmark, France, Germany, Hongkong, Italy, Japan, Neitherlands, Spain, Sweden, Switzerland, UK and US) for the period January

1970 – December 2006 and found the effect is significant in 7 countries, while a surprising September seasonal is significant in 13 countries

Gultekin (1982) finds evidence of a seasonal pattern in the stock return in most of the major industrial countries (Australia, Belgium, Canada, Denmark) in the period 1/1959 to 12/1979) The seasonality is usually manifested in a significantly larger mean return at the turn of the tax-year For most countries, this large return occurs in January

2.1.1.3 January Effect is not unique to small firm but it is more statistically significant with small firms

Thaler (1987) concludes that January effect is primary a small firm phenomenon The logic is that an equal-weighted index is a simple average of the prices of all firms listed on the NYSE Small firms have greater weight than their share of market value

Gica & Balint (2012) test Romania market from January 2003 to December 2010 and find that in the pre-crisis period (2003 to 2008), the January Effect only exists in small-cap stocks

Later on, the researchers want to find out if January effect is exclusively to small firm effect Kohers and Kohli (1991) use the data of the S&P composite index, which consist of large firm securities, and the results show that January effect was independent of the small firms effect That is, January effect was not exclusively small firm effect

10

It can be concluded that January effect is not unique to small firms but it is more common with small firms, especially for those whose prices had declined the previous years

2.1.2 Reasons for the January Effect

Most previous studies about the January Effect have indicated that some reasons for the January effect as below:

2.1.2.1 The theory of liquidity (Liquidity hypothesis): Ogden (1990) states that

cash flow tends to increase in January because at the end of year, the company pay salary and bonus to motivate employee for their contribution in company’ operation, so investors’ bonuses and salary are indirectly or directly invested on market including stock market These cash flows are then put into securities directly or indirectly through either mutual fund or pension fund, followed by an upward trend in January

2.1.2.2 Tax avoidance hypothesis: Branch (1977) points that December is the

time of tax finalization so investors tend to sell loss securities to reduce taxes on financial assets Because the tax is applied for return that happened, also investor shall sell unrealized loss stocks to reduce portfolio returns, thereby decreasing income tax When supply increases, it would put pressure on reducing the price

of securities in December However, in January, starting the new fiscal year, stock index goes up since investors tend to buy back securities to rebalance a new portfolios’ investment

2.1.2.3 Window dressing hypothesis: Haugen and Lakonisk (1988) declare that

fund managing director of the company with negative profit want to avoid financial losses by selling unrealized loss securities in fund's portfolio to better their financial statements and only retained unrealized gain securities In January, they sell unrealized gain securities and retain unrealized loss securities These transactions put pressure on stock price to decrease at the end of the year and increase in the early next year

2.1.2.4 Other Macroeconomic Factors

Schneeweis and Woolridge (1979) imply that seasonality is not

11

necessarily inefficient, and some embedded factors such as the monetary policy, risk adjustment, tax consideration, and the seasonal information difference, make the equity returns seasonal Therefore, seasonality may exist in the efficient markets

2.1.2.5 Financial Payment Systems

Ogden (1990) shows the hypothesis that the standardization of the payment system in US is a reason for cash demand by the end of each month, especially December, and investors, who have substantive cash receipts at the turn of the month will increase their demand of stocks

2.1.2.6 Traditional point of view is that good news on company’s business

operation shall be published at the end of the year Stock prices increase in January to reflect such information; therefore, return is usually higher in January

as a seasonal effect

2.2 The concept and formula of Securities Index

2.2.1 The concept of securities index

Securities index provides information for investors to analyze and evaluate securities market in general; indicating the fluctuations of securities prices and is the fundamental assessment of securities market’s performance This is very important information for listed companies, investors, state management agencies and becomes a barometer of the economy in terms of their business operation, management, investment

Securities index is a relative measure (calculated by the number of

“points”), by comparison relationship between the current value of average stock price and the basis index at the first (basis) date (usually set at 100 or 1000 at the basis date) Securities index comprises two factors, including the number of shares outstanding (weighted factor) and stock price For example, in the early period of Vietnam securities market, VN-Index is the only index representing the market value on the stock exchange market The index is calculated by weighted market value of listed shares In July 2005, HNX-Index is constructed and used

12

as a representative index for stocks listed in Hanoi Stock Exchanges (HASTC)

2.2.2 The Formula for VN – INDEX

Securities index is computed by using the Weighted average method with the Weighted factor as the number of outstanding shares listed at the time of calculation The formula of Passcher is used as below:

)Q0

x (P0

)Q1

x (P1n

1 i

i i

n

1 i

i i

Q0 : Number of outstanding share i at the basis date

n: The number of shares to calculate the index

2.2.3 The performance of VN-Index during 2000-2013

VN-Index performance in the period from 2000 to 2013 has showed the following mean return for each month of the year:

Table 1 : Average returns by months of VNI during 2000 - 2013

Month January February March April May June

Average return 5.31% -0.24% 1.94% 4.17% 0.23% 0.58% Month July August September October November December Average return -4.06% -1.61% 0.19% -2.70% -0.95% -1.01%

Source: HOSE and own analysis

Among months’ returns, January’s return has the highest value than other remaining months Therefore, we can assume that there may exist the January effect in Vietnam’s market

The following figure is monthly VN-Index Chart from 2000 to 2013:

13

Figure 3: VN-Index performance in period 2000-2013

Source: HOSE and own analysis

Looking at this VN-Index Chart, we clearly see that VN-Index’s

Fluctuation is the highest in period from 2006 to 2009 and It is known as a

“nightmare” to almost investors when VN- Index decreased about 1000 points in

period from 2008 to 2009 The period from 2009 to 2013, the VN-Index

continuously declined 200 points But it gradually is stable tendancy

Year VNI

14

CHAPTER 3- RESEARCH METHODOLOGY

3.1 Research models for January effect

3.1.1 Non-linear pattern by a Taylor expansion

As stated by Lei Gao (2005), the starting point of this analysis is the

hypothesis of an efficient market; hence, randomness of returns can be assumed Accordingly, the model states that market returns follow a geometric random walk that is the logarithmic market indices follow a random walk The first

difference, namely the market returns of stock exchange i at time t labeled r it, is stationary processes Inserting a set of dummy variables denoted d j controls for monthly effects Note that we always use July as reference month

) 2 (

11

1

it j

j ij i

of stock exchange i are jointly not significantly different from zero shall be tested

If this joint hypothesis is true then one can confirm the existence of the January

effect and the efficient market hypothesis is not true

Besides regression of the above equation, Gao (2005) tries to approximate the non-linear monthly time pattern by a Taylor expansion The first step is to

specify variable month denoted m that takes values between one and twelve

Then, the squared or cubic variable labeled 2

m , respectively, are calculated Sufficient number of powers of the variable month replaces the set of dummy variables in the above equation Note that the index i now stands for individual stocks

)3(1

it p

j

j j i

r =α +∑β +ε

=

15

3.1.2 OLS methodology by Cristina Balint and Oana Gica (2012)

The model is specified as follows:

∑= + − ++

Where ht is the variance of ϵt ,Rt represents returns on a selected index, µ,

ai and ϕ i are parameters; ϵ t is an error term, ht is the variance of ϵt , and Di are

monthly dummy variables such that Di, = 1, for the ith month and equal zero otherwise The dummy variables indicate the month of the year and i takes value from 1 to 11 (from February (1) to December (11)) In this regression, an autoregressive term is added to cope with any serial correlation, which may be caused by non-synchronous trading in stocks The test for January effect is simply a test of significance of the estimated coefficient µ

3.1.3 Time series Garch framework

As stated in Qian Sun and Wilson H.S Tong (2010)’s research, the basic GARCH (1,1) model with a January dummy is specified as follows:

)6(2

1 1

R =α +α − +α +ε εtφt−1 ≈N(0,h t)

) 7 (

3 2 1 2 1 1

3.2 Estimation of research model, description and variable measurement

3.2.1 Return calculation formula

The natural log of the relative price is calculated to produce a time series of Daily continuously compounded returns, such that:

𝑅𝑅𝑡𝑡 = ln( 𝑃𝑃𝑡𝑡

𝑃𝑃𝑡𝑡−1 ) (8)

) 5 (

1 2 1 1

0 b c feb c dec

b

16

Where:

𝑅𝑅𝑡𝑡: Monthly return in period t

𝑃𝑃𝑡𝑡: VN-Index closed every month in period t The reasons to choose logarithm returns over general return are justified by both theoretically and empirically Theoretically, logarithmic returns are analytically more tractable when linking together sub-period returns to form returns over longer intervals Empirically, logarithmic returns are more likely to

be normally distributed which is prior condition of standard statistical techniques (Strong, 1992)

3.2.2 Application of regression model to test the January effect on VN-Index

Among five models listed in the literature review, Qian Sun and Wilson H.S Tong (2010)’s regression model with a January dummy variable is used However, in this thesis I test the January effect on return, not on return’s risk or volatility, so the following equation is used:

)9(2

1 1

H : α2 >0 The January effect exists

I apply this model because it is tested to be conformed to emerging market like Vietnam securities market, such as China, India and Hongkong In addition, this model does not only use January as a dummy variable but also regress with the lag termR t−1 It is good that the lag return term is added to the mean equation

17

to filter out possible first-order serial correlation in the return series In addition, the inserting of lag term R t−1 shall make the model more confident through the increasing of R square The dummy variable for January in the mean equation is used to capture a possible January effect in the return series If the effect exists,

2

α will be positive and significance (using t-test)

3.2.3 Application of regression model to test the January effect on VN-Index’s

sub-periods

It is important to know the effect of January on different stages of the market In this research, I apply regression model to see how the impact of January effect on periods before the economic recession and during the economic recession The model is as below:

)10(2

1 1

The dummy variable for January in the mean equation is used to capture a possible January effect in the return series If the effect exists, α will be positive 2and significance (using t-test) The period before economic recession is from

2000 to 2007 The period in during economic recession is from 2008 to 2013

3.2.4 Application of regression model to test the January effect on company’s

18

market capitalization

Besides, to see further the level of January effect on different market capitalization of companies (small-cap companies, middle-cap Company and large-cap Company); the above regression model shall be applied for investment portfolio of those companies categorized by its market capitalization Each investment portfolio shall include 10 stocks of 10 companies listed in the stock market

The natural log of the relative price is calculated to produce a time series of Daily continuously compounded returns, such that:

) / )

P : Stock price at time (t – 1)

Rate of return rate on portfolio P:

∑

=

i s s

p w R R

R : Return rate of stock s

m: Total number of stocks in the portfolio

I then apply regression model to see how the impact of January effect on periods before the economic recession and during the economic recession The model is as below:

)13(

2 1 1

i t

i t

i

JAN R

Where R t i: return at month t of the portfolio i

19

The following hypothesis shall be tested:

3.3.1 Securities market data

The data to be used for empirical research is monthly returns of VN-Index from September 2000 to December 2013 The data source of securities index is from Hochiminh Stock Exchange after the adjustment of dividend and stock split

3.3.2 Company data

The securities market in Vietnam is remarkably expanded since 2007 Most companies are listed during booming period of the market: 2007-2008 In order to have enough company data for statistical analysis with the same period

of listing, the company data I collect is from January 2008 to December 2013

In Vietnam, the minimum requirement of charter capital for a listed company in Hochiminh Stock Exchange is 80 billion VND Therefore, the stocks with market capitalization below 100 billion are considered in micro-cap group Also based on market capitalization of medium and large companies in Vietnam securities market, I classify the market-cap group as below:

20

Table 2: Market capital classification

No Stock group Market capitalization

In the world, there is no specific definition for stock classification in terms

of blue chip and penny chip Blue chip is known a reputable stock with long history of business development and has solid financial position Those companies are less risky and more stable business with economic fluctuation

On the contrary, penny stock is known as stocks with lower price and lower market capitalization A typical penny stock is a stock of a company with low business scale, low liquidity and high speculation

From the above classification, portfolio investment as of 31 December

b Mid-cap portfolio

Table 4 - Mid-cap securities portfolio

STT TICKER

MKC (Bill

Source: HOSE, HNX and my own calculation

The analysis of data shall be conducted in next chapter to test the January effect on the whole period, sub periods and different market cap portfolios

22

CHAPTER 4– DATA ANALYSIS AND RESEARCH RESULTS

4.1 The January effect on VN-Index in the period 2000 – 2013

There have 161 observations of monthly returns from August 2000 to

December 2013 The number of observations with the histogram and descriptive statistics as below:

Figure 4: VN-Index histogram and descriptive statistics

Source: HOSE and own calculation

Skewness = -0.11 indicates that the distribution of monthly return is slightly skewed toward the left or the mass of the distribution is concentrated on the right of the figure The long tail points toward the low end of the distribution

or more monthly returns among observations have higher value than mean value

of 0.11%

Kurtosis > 3 means that the distribution of monthly returns has a distinct peak near the mean, decline rather rapidly and has a heavy tails with more extreme values The below table shows that most monthly returns have values in the interval: from -20% to 20%

Table 6 - Tabulation of VN-Index Return in period 2000 - 2013

Mean 0.001122 Median -0.002300 Maximum 0.325824 Minimum -0.420634 Std Dev 0.112288 Skewness -0.119642 Kurtosis 4.155294

Jarque-Bera 9.337736 Probability 0.009383

23

Sample: 2000M08 2013M12

Included observations: 161

Number of categories: 5

Cumulative Cumulative Value Count Percent Count Percent [-0.6, -0.4) 1 0.62 1 0.62 [-0.4, -0.2) 7 4.35 8 4.97 [-0.2, 0) 79 49.07 87 54.04 [0, 0.2) 66 40.99 153 95.03 [0.2, 0.4) 8 4.97 161 100.00 Total 161 100.00 161 100.00

Source: HOSE and Own calculation

To capture a positive January effect, the dummy variable for January is used By using least square method and ARMA with Eview 6.0, I regress the monthly returns with dummy variable and lag return variable as the followings:

Table 7 - Regression model for VN-Index returns in period 2000-2013

Dependent Variable: RETURN

Method: Least Squares

Sample (adjusted): 2000M09 2013M12

Included observations: 160 after adjustments

Variable Coefficient Std Error t-Statistic Prob

RETURN(-1) 0.363278 0.073034 4.974123 0.0000 DUMMY 0.056857 0.028771 1.976169 0.0499

Durbin- Watson Stat 1.892

Source: HOSE and Own calculation

DW = 1.892 implies that this model is suitable for testing the January effect in Vietnam Securities market

The coefficient of dummy variable is positive and p-value is less than 5%,

24