Whether momentum or contrarian phenomenon exist in vietnam stock market

The screen of the Analyzing Stock Price Data after stocks are ranked in descending order on the basis of their average monthly returns………...26 Figure 4.3: The screen of Stock Grouping

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

-o0o -

TA THU TIN

WHETHER MOMENTUM OR CONTRARIAN PHENOMENON EXIST IN

VIETNAM STOCK MARKET MAJOR: FINANCE – BANKING

MAJOR CODE: 60.31.12

MASTER THESIS

ADVISOR: Ph.D TRAN PHUONG NGOC THAO

HO CHI MINH CITY, 2011

My sincere thank goes to Nguyen Hiep Phat, my colleague at Au Viet Securities, he spent

a lot of time to help me make a software program to process raw data in this thesis

Finally, I am thankful to my family for giving me facilitation to complete my thesis

ABSTRACT

This thesis investigates whether momentum or contrarian phenomenon exist on Vietnamese Stock Market over the period from January 2005 to June 2011 We employ the famous methodology by Jegadeesh and Titman (1993) to calculate the profit of momentum and contrarian strategies which were built base on the historical return of 424 stocks listed on Ho Chi Minh Stock Exchange and Ha Noi Stock Exchange

We found that all 16 trading contrarian strategies always make abnormal profit with statistical significance at the level of 10% The most profitable contrarian strategy with portfolio based on 6 month formation and 3 month holding has a average monthly return

of 2,829% (equivalent to annually return of 33,95%) with significance level of 2%

Our research demonstrates that the abnormal profit on trading contrarian strategy can not

be accounted for by beta-risk as well as market size But we found a evidence of P/B ratio explaining contrarian phenomenon on Vietnamese Stock Market

Key words: Momentum; Contrarian strategies

TABLE OF CONTENTS

ACKNOWLEDGEMENT……… i

ABSTRACT……… ii

TABLE OF CONTENTS……… iii

LIST OF FIGURES……… vi

LIST OF TABLES……… vii

ABBREVIATIONS……… viii

1 Introduction……… 1

1.1. Overview of Momentum and Contrarian strategies……… 1

1.2. Research Objective……… 2

1.3. Research Methodology and Scope……… 3

1.4. Thesis Structure……… 4

1.5. Vietnamese Stock Market……… 5

2 Literature Review……… 7

2.1. Efficient Market Hypothesis……… 7

2.2. Momentum Strategy……… 8

2.3. Contrarian Strategy……… 15

3 Data Collection and Research Method……… 18

3.1. Data Collection……… 18

3.1.1 Stock Prices……….……… 18

3.1.2 Adjusted Stock Prices……… 19

3.1.3 P/B ratio……… 21

3.1.4 Market Capitalization……… 22

3.2. Research Method……… 22

4 Empirical Result……… 24

4.1. Raw Data Processing……… 24

4.2. Empirical Result……… 28

4.3. Why does the contrarian phenomenon exist in Vietnam stock market? 32

4.4. Some factors may account for the contrarian phenomenon in Vietnam Stock Market……… 35

4.4.1 Market Risk……… 36

4.4.2 Firm Size……… 39

4.4.3 Price to Book……… 40

5 Conclusion……… 42

5.1. Main findings……… 42

5.2. Implications of Research……… 43

5.3. Limitations of Research……… 44

REFERCENCES……… 45

APPENDIX……… 48

Table A1: List of 424 investigated stocks………48

Table B1-Table B16: The average monthly return of Winner and Loser Portfolios in 16 strategies……… 59

Table C1-Table C16: The average monthly return of Loser portfolio compare to the one of Winner portfolio in 16 strategies……… 75

Table D11: Estimation of Beta of Winner portfolio in J=3/K=3 strategy……… 83

Table D12: Estimation of Beta of Loser portfolio in J=3/K=3 strategy……… 84

Table D21: Estimation of Beta of Winner portfolio in J=6/K=3 strategy……… 85

Table D22: Estimation of Beta of Loser portfolio in J=6/K=3 strategy……….……… 86

Table D31: Estimation of Beta of Winner portfolio in J=9/K=3 strategy……….87

Table D32: Estimation of Beta of Loser portfolio in J=9/K=3 strategy………88

Table D41: Estimation of Beta of Winner portfolio in J=12/K=3 strategy……… 89

Table D42: Estimation of Beta of Loser portfolio in J=12/K=3 strategy……….90

Table E1: Average Market Capitalisation of Loser portfolio compare to Market Capitalisation of Winner portfolio in J=3/K=3 strategy……… 91

Table E2: Average Market Capitalisation of Loser portfolio compare to Market Capitalisation of Winner portfolio in J=6/K=3 strategy……… 91

Table E3: Average Market Capitalisation of Loser portfolio compare to Market

Capitalisation of Winner portfolio in J=9/K=3 strategy……… 92

Table E4: Average Market Capitalisation of Loser portfolio compare to Market

Capitalisation of Winner portfolio in J=12/K=3 strategy……… 92

Table F1: The average P/B ratio of Loser portfolio compare to

the average P/B ratio of Winner portfolio in J=3/K=3 strategy……… 93

Table F2: The average P/B ratio of Loser portfolio compare to

the average P/B ratio of Winner portfolio in J=6/K=3 strategy……… 93

Table F3: The average P/B ratio of Loser portfolio compare to

the average P/B ratio of Winner portfolio in J=9/K=3 strategy……….94

Table F4: The average P/B ratio of Loser portfolio compare to

the average P/B ratio of Winner portfolio in J=12/K=3 strategy……… 94

LIST OF FIGURES

Figure 1.1 VN-Index Chart over the period from July 2000 to September 2011………….6 Figure 3.1 Formation and Holding periods in two strategies……… 23 Figure 4.1 The screen of the Analyzing Stock Price Data program

after importing data from Excel file……….25 Figure 4.2 The screen of the Analyzing Stock Price Data after stocks are

ranked in descending order on the basis of their average monthly returns……… 26 Figure 4.3: The screen of Stock Grouping program shows the Winner

and Loser portfolios, and their average returns……… 27

LIST OF TABLES

Table 3.1 Adjusting price of KDC share……….… 20

Table 3.2 Adjusting price of OPC share……….… 21

Table 4.1 The average return of Winner and Loser Portfolios and their difference in J=3/K=3 Strategy (Formation Period: 3 months; Holding Period: 3 months)…… ……28

Table 4.2 Summary of the average monthly return of loser and winner portfolio;

and their differences (profitability of contrarian strategies) for 16 strategies

over the period from 01/2005 to 06/2011……….… 30 Table 4.3 Monthly and annually profitability of 16 contrarian strategies

are ranking in descending and their significances……… 31

Table 4.4 Beta coefficient after perform regression……… 38

Table 4.5 The comparison in average market capitalization between

loser and winner portfolios and their differences……… 39

Table 4.6 The comparison in average P/B ratio between loser and winner portfolios…41

ABBREVIATIONS

EMH Efficient Market Hypothesis

CHAPTER 1: INTRODUCTION

1.1 Overview of Momentum and Contrarian Strategies

In 1970, Efficient Market Hypothesis (EMH) developed by Professor Eugene Fama proclaimed that in the efficient market no one could consistently beat the market and stock prices follow a random walk Thus, future prices of stocks could not be predicted from their past prices, it means that the abnormal return from trading should be zero However, a lot of investors and researchers have doubts about the efficient market hypothesis both empirically and theoretically They always try to find some abnormal returns to prove the inefficiency of the markets

Consequently, forecasting the price movements in stock markets has become a major challenge for investors, brokers and speculators Studying the movement of stock prices become one of the most attractive fields of research due to its commercial applications and benefits it offers Recently, there are many researchers and traders have studied stock price predictions such as Fundamental Analysis, Technical Analysis, CANSLIM, etc…

And one of the most attractive trading strategies is momentum (and contrarian) strategy The momentum strategy appeared firstly in the 1960s However, it became widely known only in the early 1990s after Narasimham Narasimhan Jegadeesh and Sheridan Titman published their study Momentum and contrarian strategies are two opposite investment strategies which use historical price/return data in order to forecast the future development of stock performance to make excess returns Momentum investing strategy, also sometimes known as “Trend following”, believes that stocks which have good performance in the past will keep doing so in the future, it buys (go long) stocks that have outperformed in the recent past, and short sell (go short) those that have underperformed over the same period In contrast, a contrarian strategy believes that stocks which have good historical performance will be bad in the future, so it suggests short selling past winning stocks and buying past losing stock The contrarian strategy was introduced first

by DeBondt and Thaler in 1985, they report that long term past losers outperform long term past winners over the subsequent three to five years

After the appearance of this strategy, a lot of studies were carried out to prove or deny abnormal returns However, most momentum strategy studies have concentrated exclusively on the US market and only a few authors have touched on non-US stock exchanges Among non-US studies, the majority of papers investigated momentum in developed markets rather than in emerging markets As momentum profits may be explained by market inefficiency, we hypothesize that underdeveloped stock exchanges may show higher momentum due to their lower efficiency level

In this thesis we will investigate two types of investment strategies: momentum and contrarian strategy in the Vietnamese stock market

1.2 Research Objective

In this thesis we test whether the momentum (or contrarian) strategy make abnormal over time horizons by examining portfolios which are formed on the historical returns, with the top decile of the ranked stocks labelled the winner portfolio and the bottom decile labelled the loser portfolio

The Efficient Market Hypothesis predicts that these winner/loser portfolios will yield zero profits, therefore if we find out that the momentum or contrarian momentum exist on Vietnam Stock Market, we have an evidence prove that Vietnamese Stock Market is not a weak form of Efficient Market Hypothesis

This study has three main following objectives:

1 Investigating whether the momentum or contrarian phenomenon exist on the Vietnamese Stock Market?

2 Determining factors account for the momentum or contrarian phenomenon

Research Questions

Three above research problems lead to the following research questions:

1 Based on historical stock price data in Vietnamese stock market, would momentum (or contrarian) strategy make abnormal profit?

2 Why does Vietnamese stock market have momentum/contrarian phenomenon?

3 Which factors explain the abnormal return yielded by momentum/contrarian strategy?

1.3 Research Methodology and Scope

The object of this research is market prices of 424 listed stocks in Ho Chi Minh Stock Exchange (HOSE) and Ha Noi Stock Exchange over the period from January 2005 to June 2011

In this thesis, we use the famous method which was proposed by Narasimhan Jegadeesh and Sheridan Titman in 1993 to test the profitability of momentum (contrarian) strategies

In each month we form a winner portfolio with 10% top stocks with highest performance and loser portfolio with 10% top bottom stocks with lowest performance Then we make

a comparison the average past return of stocks in the winner portfolio with these in the loser portfolio in next K months

In this thesis we use two software programs to analysis data: Analyzing Stock Price Data and SPSS Analyzing Stock Price Data is a small software composed by Nguyen Hiep Phat an Information Technology Engineer working at Au Viet Securities Corporation This software ranked stocks ascending based on the average past returns Next, it build winner and loser portfolio The winner portfolios consist of the top decile which comprises the stocks with the highest performance in the previous J months, and the loser portfolios consist of the bottom decile which comprises the stocks with the lowest performance Then it makes a comparison between the average past return of stocks in the winner portfolio and these in the loser portfolio

We use SPSS software to perform the paired sample t-test in to make comparison between the average returns, market sizes and P/B ratios of loser portfolio and these of winner portfolio We also use ordinary least square regression in SPSS to estimate the market risk (beta) of loser portfolio and winner portfolio

1.4 Thesis Structure:

The thesis goes through 5 following chapters

Introduction: This chapter introduces research background of the study, research problems, research objectives, data and research methodology Furthermore, this chapter gives a little introduction about Vietnamese Stock Market

Literature review: This chapter will clarify what has already been observed and documented in the area of momentum and contrarian phenomena The literature review

is divided into three main parts in order to consistently illustrate how the previous researchs are relevant to our study Firstly, we give an overview of the Efficient Market Hypothesis Secondly, we investigate the momentum strategy- buying past winners and selling past losers And finally we examine the opposite trading strategy to momentum, which is widely called contrarian strategy-buying past losers and selling past winners

Data and Methodology: This chapter divided two part The data part show how can we gather market price, market capitalisation and book value per share of stocks and the way

we adjust stock’s market price used in the research Then we present the method we use

to calculate the profit made by using the momentum (or contrarian) strategies

Empirical Result: Thi chapter investigates whether the momentum (contrarian) strategy has existed on the Vietnamese stock market over the period from January 2005 to June

2011 Furthermore, various robustness tests are conducted

Conclusion: The conclusion sums up on the entire thesis, thereby answering the initial problem statements, and also giving some recommendations and limitations of this study

1.5 Vietnam Stock Market

To perform the industrialization and modernization process of the country, Vietnam has been requiring a huge capital Consequently, building stock market in Vietnam had become an urgent task to mobilize midterm and long term capital Furthermore, the development of stock market helps Vietnam equitize and restructure its state-owned enterprises

On July 10th 1998, Prime Minister Phan Van Khai signed a decision to set up two securities trading centers at Hanoi and Ho Chi Minh City The first session was launched

on July 28th 2000 at the Ho Chi Minh City Securities Trading Center And five years later, in March 2005 the second Hanoi stock trading center (Hastc) was established

Currently, there are three stock exchanges operating in Vietnam, the Ho Chi Minh Stock Exchange (HOSE) in Ho Chi Minh City, Ha Noi Stock Exchange (HNX) and UpCom Market (Upcom stands for Unlisted Public Company Market) in Ha Noi

As of June 2010, there are 622 listed companies and 4 listed investment funds were traded on the 3 Vietnamese Stock Exchanges, with a total market capitalization of VND746.76 trillion (US$39.1 billion- accounted for 40% of GDP)

Thanks to the growth of the stock market, the number of securities companies increased rapidly during the period from 2007 to 2008 Currently, there are 101 securities companies registering to be members of HOSE and HNX with total registered capital of VND 33,340 billion Most of securities companies have been granted licenses covering all kinds of businesses: brokerage, dealing on own accounts, underwriting and investment advisory

A rapid growth of Vietnamese stock market has attracted many local and foreign investors, individuals as well as institutions At the end of August 2011, the number of investor accounts opened in member securities companies was more than one million in which foreign investors made up over 15,350 accounts, over 1,000 of which belong to institutional investors, account for 20% market size

From July 2003, foreign investors were allowed to buy up to 30 percent of the chartered capital of a listed company This figure was then raised by to 49 percent at the end of the third quarter of 2005

The movement of Vietnamese stock market can be observed clearly via VN-Index’s changes Because of the highly speculative nature of Vietnamese stock market, there was

a considerable fluctuation in VN-Index throughout the period from 2005-2011 From 307.5 points in late 2005, VN-Index reach the peak 1,173 points on March 12th, 2007, then it slipped to 235 points in February 2009 And now VN-Index level off around 400 points Bellow image illustrates the movement of VN-Index from July 2000 to September

2011

Figure 1.1: VN-Index Chart over the period from July 2000 to September 2011

CHAPTER 2: LITERATURE REVIEW

This chapter literature review is divided into three main parts in order to consistently illustrate how the previous studies are relevant to our research Firstly, we give an overview of basic knowledge of Efficient Market Hypothesis, CAPM Model In the second part, we concentrate to investigate preceding studies about momentum strategy – buying past winners and short selling past losers And in the third part we investigate the opposite trading strategy to momentum – buying past losers and short selling past winners, this strategy is called contrarian strategy

2.1 Efficient Market Hypothesis

One of the most pivotal concepts in finance is the Efficient Market Hypothesis, it was introduced by Eugene Fama in 1970 Fama defined that an efficient market as one in which all available information is reflected in the market price In his publication, Fama finds statistically highly significant support that stock prices follow a random walk It is impossible to consistently outperform the markets by using any historical information the market already knows It means that the abnormal return from trading should be zero

The efficient market hypothesis is closely related to the random walk hypothesis Therefore, it can be said that in efficient market stock prices follow a random walk, and thus future prices cannot be predicted from their past prices There are three levels of the efficient market: weak, semi-strong, and strong

Weak form of efficient market hypothesis: Which says that all past publicly information contained in past stock price movement is reflected in current market price, nothing is overvalued or undervalued The price changes follow a random walk process It is impossible to predict future prices by analyzing historical prices Hence, technical analysis expert will not be able to make exceed return, reading charts or trend analysis by technical analysts is just time wasted

Semi strong form of efficient market hypothesis: It is implied that stock prices reflect all publicly available information about the company’s prospects such as anticipated earnings, earning announcements, stock issues, the company’s credit rating, the success

of new products… and stock prices will adjust immediately to reflect new public information So, excess returns cannot be earned by trading on that information Semi strong form of efficient market implies that neither fundamental analysis nor technical analysis techniques will be able to produce exceed returns

Strong form of efficient market hypothesis: In this form, all information relevant to the company is considered as available, even hidden or insider information And, stock prices reflect all these information No investors can earn excess returns as these information have already been priced There is evidence for and against the weak and semi-strong EMH, while there is powerful evidence against strong EMH

However, a dispute about market efficiency is inevitable A lot of investors and researchers have doubts about the efficient market hypothesis both empirically and theoretically So they always try to find some trading strategies or methods that yield a abnormal returns than the market with equivalent risk to prove the inefficiency of the markets The financial industry as well, in particular within the asset management, takes a big interest in finding and exploiting anomalies based on market inefficiency

This part will focus on the concept of momentum phenomenon and the development of momentum research Momentum is the phenomenon that prices of rising assets tend to keep rising in the future or that past winners will yield a higher return than past losers The following, also presents the empirical findings from a number of studies of momentum phenomenon and momentum strategies

Momentum investing strategy, also sometimes known as “Trend following”, basically is an investing philosophy which is based on a past trend of stock price Specifically, it is believed that stocks which have good performance in the past will keep doing so in the near future,

then an investment approach that buys stocks that have high returns in recent time and short sells stocks that have poor returns will outperform the market This is a investing strategy based on the underlying belief that stock price trends generally continue for a long period of time

The first debut of momentum in stock prices had been conducted by Alexander in 1961

He examined the behavior of stock prices in speculative bubbles and found that “there are trends in stock market prices, once the “move” is “taken” The next remarkable paper was published by Levy in 1967 who documented that the strategy of buying stocks priced significantly higher than their average prices over the past 27 weeks generated abnormal returns

A next achievement in momentum phenomenon was investigated and proved thoroughly

by Narasimhan Jegadeesh and Titman (1993) They published a significant paper

“Returns to Buying Winners and Selling Losers: Implications for Stock Market Efficiency” in which they introduced the momentum trading strategy into financial literature They analyze the data of American stock prices listed on the NYSE and the AMEX stocks during the sample period from 1965 to 1989 They concluded that the strategy based on forming a portfolio while buying the top 10% of stocks with past highest returns and short selling the bottom 10% of stocks with past lowest return will be profitable over medium-term (between 3 and 12 months) In this study all returns are statistically significant, except for the 3M/3M strategy (3 months formation and 3 months holding) that does not skip a week between the formation and the holding periods The most successful momentum strategy in this study is the 12M/3M strategy, which make a return of 1.31% per month (t=3.74) when there is no time lag between the formation and the holding periods and 1.49% per month (t=4.28) when there is a 1-week lag between the formation and the holding periods The worst strategy is the 3M/3M strategy, which had the return of 0.32% (t=1.10) without one week lag between the formation and the holding periods, and 0.73% (t=2.61) with a 1-week lag The 6M/6M strategy with no time lag between the formation and the holding periods had an average monthly momentum return of 0.95% (t=3.07) In general, it seems that strategies with long

formation periods of 9 or 12 months and short holding periods of 3 or 6 months perform somewhat better than the remaining strategies

Louis K.C Chan, Narasimhan Narasimhan Jegadeesh, and Josef Lakonishok (1996) examine the momentum strategy with 6 month formation period over a sample period from 1977 to 1993 using primarily stocks listed on the NYSE, AMEX, and NASDAQ Their results show a momentum return of 8.8% with the 6 month holding period and in years 1, 2 and 3, the momentum returns are 15.4%, -0.6% and 1.2%, respectively

In 1998, Conrad and Kaul investigated the momentum strategies which have the length of the formation and the holding periods spread between 1 week and 36 months To make a comparison with Narasimhan Narasimhan Jegadeesh and Titman’ study (1993), Conrad and Kaul initially investigate stocks listed on the NYSE and AMEX during the period from 1962 to 1989 They find that, with the exception of the 1-week/1-week strategy, all other strategies show profitable up to and including the 18M/18M strategy Therefore, this results confirm the momentum effect documented by Narasimhan Jegadeesh and Titman

In 2001, Narasimhan Jegadeesh and Titman extend the 6M/6M strategy from their original study with eight additional years Using data over the 1990 to 1998 sample period, they find that the momentum strategy continues to be profitable with an average monthly return of 1.39% (t=4.71) over the 8-year period They find that momentum effect exist in both small and large capitalization stocks, but the momentum phenomenon in the small capitalization stocks is stronger than large capitalization stocks Basically, the findings in Narasimhan Jegadeesh and Titman’s second article appear to be very similar

to those in their original study

Lee and Swaminathan (2000) test the momentum effect over the 1965 to 1995 sample period using all firms listed on the New York Stock Exchange (NYSE) and American Stock Exchange (AMEX) Lee and Swaminathan exclude National Association of Securities Dealers Automated Quotation System (NASDAQ) stocks from their sample, as

NASDAQ firms tend to be smaller and thus more difficult to trade in momentum strategies They investigate 16 different strategies and find the returns of all zero-cost portfolios to be positive and statistically significant To be similar to study obtained by Narasimhan Jegadeesh and Titman (1993) the 12M/3M strategy turns out to be the most successful with an average monthly return of 1.54% (t=5.63) over the sample period, whereas the worst performing strategy turns out to be the 3M/3M strategy with an average monthly return of 0.66% (t=3.06) Hence, also here the general pattern seems to

be that strategies with relatively long formation periods and short holding periods are to prefer

Some finding of momentum phenomenon from European market

In 1998, Rouwenhost published his study in which he examined the profitability of momentum strategies by using sample data of 2,190 companies from across 12 different European Markets over the period from 1980 to 1995 Geert Rouwenhorst finds that all returns of 16 J/K strategies that do not skip time between the formation and the holding periods are positive and statistically significant at the 5% level The average monthly returns range from 0.70% (t=2.59) using the 3M/3M strategy to 1.35% (t=3.97) using the 12M/3M strategy

Geert Rouwenhorst also found out the best and the worst performing strategies are the same as Narasimhan Jegadeesh and Titman (1993) investigated on the U.S market In this study Geert Rouwenhorst discover that momentum strategies with long formation periods of 9 or 12 months and short holding periods of 3 or 6 months tend to perform better than the remaining strategies

When comparing the returns of momentum strategies in individual countries, Geert Rouwenhorst choose the 6M/6M strategy which does not skip time between the formation and the holding periods as a representative strategy He finds that, with the exception of Sweden, all the country-specific returns are statistically significant The strongest momentum effect is found in Spain, followed by the Netherlands, Belgium, and Denmark

He also finds that the momentum phenomenon seem happen to small capitalization, Geert Rouwenhorst suggests on the basis of his findings that return continuation is stronger in smaller firms than larger firms in the sample However, he finds that, although there is a negative relation between market capitalization and the profitability of the momentum strategies, past winner stocks tend to significantly outperform than past loser stocks regardless of the market size of the stocks This result is consistent with the one found by Narasimhan Jegadeesh and Titman (1993) on the U.S market

Mark T Hon and Ian Tonks (2002) tested the profitability of momentum trading strategies in the UK stock market by using historical returns of the largest set of individual securities in the UK stock market over the period from 1955-1996 They used strategies with 3, 6, 9, 12, 15, 18, 21 or 24-month formation periods and holding periods

of 3, 6, 9, 12, 15, 18, 21 or 24-months The results show a total of 24 trading strategies that are positive and statistically significant at the level of at most 10% The most profitable strategy is the 12M/6M momentum strategy with an annualized return of 16.2%

Their study show that returns on momentum strategies cannot be accounted for by a simple adjustment for beta-risk They also find that, there are some evidences of a size effect in the UK stock market, this phenomenon could not explain the abnormal profit of momentum effect

In 2002, Dijk and Huibers (2002) replicate the methodology of Geert Rouwenhorst (1998) for their analysis of price momentum on stock data from 15 European countries in the period 1987-1999 They used strategies with 12-month formation periods and holding periods of 1, 3, 6, or 12 months Their findings are consistent with Geert Rouwenhorst’s (1998) and document that momentum strategies are profitable for all the examined holding periods, also after correction for stock-related risk, book-to-market and size effects

Markus Glaser and Martin Weber (2003) also analyze the momentum phenomenon in German Stock market by using the data comprised of 446 listed on the Frankfurt Stock in

the period 1988-2001 Instead of ranking 10 portfolios by stock returns as Narasimhan Jegadeesh and Titman’s methodology, Markus Glaser and Martin Weber use only 5 portfolios to rank the stocks on the basis of their returns Markus Glaser and Martin Weber documents that stocks with a high return will tend to have a higher price momentum effect, than stocks with a low return Although they find momentum strategies to be profitable, the turnover factor has no predictive power among stocks with

a large market capitalization according to their empirical analysis This entails that momentum strategy should focus on small capitalization stocks, which are more illiquid and therefore also costly to trade

Bird and Whitaker (2003) applied momentum strategies to seven major European markets over the period from 1990 to 2002 They used these strategies with 6 or 12-month formation periods and the holding periods of 1, 3, 6, 9, 12, 24, 36, or 48 months For the combined equally-weighted portfolios, Bird and Whitaker find that, with the 6-month formation period and the holding periods of up to 9 months, the momentum strategy make a return of more than 7% In case of the 12-month formation period, the optimal holding period is less than 6 months, with the return around 4% for the 6-month holding period Both cases of 6 and 12 month formation period, the greatest and most significant returns are found for holding periods of 3 months and 1 month

Bird and Whitaker also find that performance of the momentum strategies becomes negative over longer holding periods They find the strategy momentum with holding periods beyond 24 months and 6-month formation period are negative, and the same result to the strategy using holding periods beyond 12 months and 12-month formation period

Finally, Bird and Whitaker investigate the profitability of momentum strategy in each individual country They use a formation period of 6 months and holding periods of 6, 9,

or 12 months; and conclude these momentum strategies are profitable in all 7 countries for all holding periods; however, with a relatively low return on the French and Spanish markets

Dirk Schiereck (1999) examined the momentum effects of the German stock by alayzing all stocks listed on the Frankfurt stock exchange in the period from 1961 until 1991 The set of the formation periods includes 1, 3, 6, and 12 month formation period His results are very similar to the ones that Geert Rouwenhorst (1998); Narasimhan Jegadeesh and Sheridan Titman (1993) reported The abnormal return of zero cost portfolios with 12 month holding period are 1.49% (t=6.35), 5.52% (t=5.57), 8.07% (t=4.95), and 5.21% (t=1.87), respectively

Asian Markets

In 2000, Chui, Sheridan Titman, and Wei investigate the momentum effects on 8 Asian markets (Hong Kong, Indonesia, Japan, Korea, Malaysia, Singapore, Taiwan, and Thailand) by using a 6M/6M strategy during the period from 1976 to 2000 The results showed that the momentum effect is present in all of the Asian countries, except Korea and Indonesia, but that it is generally weak; and statistically significant in Hong Kong only When they combine stocks from all eight countries into one aggregate sample, the momentum effect still weak and to be statistically insignificant with an average monthly momentum return of 0.38% However, when excluding Japan market from aggregate sample, the momentum effect is quite strong and statistically significant In general, Chui

et al find the momentum effect to be relatively stronger for firms with small market capitalizations, low book t market values, and high turnover ratios

Kevin R Foster, Ali Kharazi (2006) studies contrarian and momentum effect on Iran Stock Market by using data of stock prices over the sample period 1997 - 2002 They apply the method of Jegadeesh and Titman (2001) and find that there is evidence of

“momentum” where past high ruturn stocks have above average return over an intermediate (3 - 12 months) horizon These result show that the stocks in the top decile tended to outperform the stocks in the lowest decile, by 3 – 11%, depending on the formation and holding period The smallest outperformance is for the 3M/3M strategy with return of 3.07% The two largest strategies are 9M/6M and 6M/9M strategies which outperform by 11.07% and 11.05%, respectively

2.3 Contrarian Strategy

The opposite of Momentum strategy is Contrarian strategy This kind of strategy is based

on the pivotal idea that investors in the stock market tend to overreact to information, implying that stock prices may also overreact Overreaction occurs according to Shleifer (2000, pp 114–121) when people overreact to unexpected new information, that is, they are willing to pay an unjustified high price for a security after the arrival of good news, respectively they are only willing to sell an unjustified low price after bad news It means that winner stocks tend to be overvalued while losers stocks are undervalued Therefore, contrarian investor can exploit this mentality to obtain financial gains when stock prices revert to their intrinsic values

Contrarian strategy literature became widely popular in the late 1980s – early 1990s The earliest and most influential evidence of contrarian phenomenon is introduced by Werner

De Bondt and Thaler (1985) who find that contrarian strategy could generate abnormal returns

De Bondt and Thaler found that a long term contrarian strategy, which buy losers and sell winners based on a performance during the 3 to 5 previous years of formation period, will generate positive returns of nearly 8% per year in the next 3 to 5 year holding periods However, many researchers attribute the performance of this contrarian strategy to investor behavior and result of Bondt and Thaler could be explained by the systematic risk (beta) of their Loser portfolios Werner De Bondt and Thaler (1985) note that past performance can serve as a proxy for investor sentiment, and since prices are initially biased either by excessive optimism or pessimism, prior losers would make more attractive investments than prior winners over the long term Werner De Bondt and Thaler’s argument is consistent with the hypothesis of long term over reaction by investors to information—a hypothesis documented in several other markets (e.g., Gunaratne and Yonesawa (1997) in Japan, Dirk Schiereck, Werner De Bondt, and Martin Weber (1999) in Germany)

Narasimhan Jegadeesh and Lehmann (1990) reports that a contrarian strategy, based on information from the previous month, yields statistically significant abnormal returns of 1.99% per month (23.88% per year) over the period 1934-1987 in United State market However, other papers argue that their results can hardly be attributed to an overreaction effect, but are rather caused by a delayed stock price reaction to common factors (Lo & MacKinlay, 1990), the presence of short-term price pressure or a lack of market liquidity (Narasimhan Jegadeesh & Sheridan Titman, 1991)

Clare and Thomas (1995), using a sample of 1000 stocks over the period 1995 to 1990 in the UK stock market, find similar result with De Bondt and Thaler That is, loser portfolio outperforms previous winners over a two year period by statistically significant 1.7% per annum They also find the losers often tend to be small, and the limited overreaction effect is probably due to the size effect

Qiwei Chen, Ying Jiang, Yuan Li (2010) investigate contrarian strategies in the Chinese market, which include all domestic stocks (A shares) listed on both Shanghai and Shenzhen Stock Exchange over the period from 1995-2010 They find statistically significant short term contrarian profit of around 0.8% per month, with 1 to 4 month holding period based on previous 1 to 2 month formation period The contrarian strategy generates higher profit in the period of bear market than in the period of bull market, this suggests that a contrarian strategy should be used as when the market is in decline

Tibebe Assefa and Omar Esqueda (2010) investigate the large-cap stocks (their current market capitalization of five billion U.S dollars or more) in the United State over the period from January 1976 to December 2008, and find that both the loser and the winner portfolios gain in the test periods, but the loser gained more (29.2%) than the winner at

36 months after portfolio formation In order to explain the differential returns between the winner and loser portfolios, they further estimate the CAPM for both winner and loser portfolios The results indicate that the losers, on average, have lower beta (β=0.98) than the winners (β=1.29) This result indicates the higher returns of the winner portfolios is

the result of a higher risk associated with them This suggests that CAPM alone is not enough to explain the abnormal return between the winner and the loser portfolios

To sum up, even though the findings of the papers mentioned above concerning contrarian strategy are debated by other researchers, they find two possible time spans for when abnormal profits may exist – very short term (over a week to a month) and very long term (over 3 to 5-year period)

Empirical evidence suggests that these two strategies mutually co-exist, since the contrarian strategy is supported for very short term holding period (usually around one month) and long term period (usually more than 36 months), while the momentum strategy is profitable in short to medium horizons Subsequent studies have demonstrated that the profitability of both contrarian and momentum strategies are international (e.g., Griffin et al, 2003, Clare and Thomas, 1995, Chui et al, 2005, Hameed and Kusnadi, 2002) Although there are sufficient supportive evidence for both strategies, the source and interpretations of the profits is a subject of much debate

CHAPTER 3: DATA COLLECTION AND

The data for this study includes 424 stocks listed on the Ho Chi Minh Stock Exchange (HOSE) and the Ha Noi Stock Exchange (HNX) and these stocks satisfy the required conditions with having at least 18 trading months The sample covers the time period from the January 2005 to the end of June 2011 There are a total of currently listed stocks

on HOSE as well as 250 listed stocks on HNX

In this section will demonstrate how we collect and calculate market price of stocks, adjusted stock prices, number of outstanding shares of stocks, book value per share of stocks, market capitalization of stocks

3.1.1 Stock Prices

In this thesis, the data of stock prices are mainly gathered from Phu Toan Investment Research Corporation and Au Viet Securities Corporation The prices were the closing price at the end of the month All listed stocks having at least 18 trading months are included in this study regardless of their low market capitalizations, low liquidities or small prices

3.1.2 Adjusted Stock Prices

In this section we will describe how I adjust stock price after firm’s announcement

If a company pay cash dividend or stock dividend, a comparison between historical stock price to present price day would not be accurately reflected For this reason, we must adjust stock price to compare them and calculate stock return

When we adjust stock price of a listed company on Ho Chi Minh Stock Exchange, we use price of the stock at the end of every trading session While calculating the adjusted price

of stock listed on Ha Noi Stock Exchange, we use the weighted price of stock in every trading session

All adjusted closing price of stocks used to calculation in this thesis are closed prices at the end of a month The monthly closed price is used to compare a stock’s performance

in a period of time

During the period time of trading, many things which can affect a stock price, a lot of news relating to the operations of a company, besides there are some distributions of company also influence stock price These distribution events are cash dividends, stock dividends or stock splits When distributions are executed, the closing prices of all previous trading days are adjusted according to an appropriate proportion

Following show the method we use to adjust price of stock after firm’s announcements

For cash dividends, the value of the dividend (D) is subtracted from the last closing price

of the stock

For example, Kinh Do Corporation (KDC) announces that it pay a dividend of VND 1.200 per share, ex-right date is on the August 24th 2011 The closing price of KDC share is VND 36.000 on the Agust 23th 2011 So, the adjusted closing price for the stock

on 23/08/2011 would then be VND 34.800 (VND 36.000-VND 1.200) And doing the same thing for all previous prices of KDC share

This could be presented clearly in the bellow table 1

Table 3.1: Adjusting price of KDC share

Closing price …… 33.900 34.300 36.000 36.000 35.800

Adjusted closing price …… 32.700 33.100 34.800 34.800 35.800

For stock dividends (or for bonus shares) If ABC Corporation announces that a exercise ratio (m:n) stock dividend instead of a cash dividend, the adjusted closing price calculation will change A ratio (m:n) stock dividend means that for every (m) shares an investor owns, he will receive (n) shares, and the total shares he will own is equal to (m+n) shares when the distribution is executed In this case, the adjusted closing price calculation will be:



For stock dividends (or for bonus shares) and cash dividends If XYZ Corporation announces a ratio (m:n) bonus shares, and investor will receive amount (D) cash dividend for every share he already owns This time the calculation of adjusted closing price calculation will be



For example OPC Pharmaceutical Joint-Stock Company (OPC) announces that it pay a dividend of VND 1.000 per share, and issue bonus shares to the existing shareholders with exercise ratio 2:1, and the ex-right date is on the August 12th 2011 The closing price

of OPC share is VND 35.100 on the Agust 11th 2011 So, the adjusted closing price for

the stock on 11/08/2011 would then be VND 22.700 ((35.100-1.000)*2)/(2+1)) And

doing the same thing for all presious prices of OPC share

Table 3.2: Adjusting price of OPC share

Closing price ……… 37.500 37.000 35.700 35.100 23.500 Cash dividend (D) ……… 1.000 1.000 1.000 1.000 1.000

Adjusted closing price ……… 24.300 24.000 23.100 22.700 23.500

We have just considered some most common events that could affect a stock's closing price However, there are more complicated actions, such as issuing new shares, issuing bonds or convertible bonds And the general technique which could be applicable to all case is that adjusted closing price will be equal to total amount payment of investor divided by total number of shares investor will receive

3.1.3 P/B ratio

P/B ratio is used to compare a market price of stock to its book value In this thesis P/B ratio is calculated by dividing the closing price of the stock by the latest quarter's book value of the stock

The book value of stock is calculated by dividing the equity capital by the number of outstanding share of that stock The number of outstanding share of stock is calculated by dividing the chartered capital by 10,000

The book value of stocks are collected from three main sources: their financial reports published quarterly, Stoxplus Corporation and Au Viet Securities Corporation

3.1.4 Market Capitalization

Market capitalization is a measurement of the worth of a company It represents the aggregate value of a stock interest that shareholders hold in a company In this thesis, market capitalization is calculated by multiplying a current number of outstanding shares

of a company by the adjusted market price of stock

It is similar to the book value, the number of outstanding shares are collected from three main sources: their financial reports published quarterly, Stoxplus Corporation and Au Viet Securities Corporation

3.2 Research Method

This section will present the most commonly used method when testing the profitability

of momentum strategies It should be noticed that this method is very similar to the one used by Narasimhan Jegadeesh and Sheridan Titman (1993), as they provided the pioneering academic work on momentum strategies, and other researchers to a large extent have adopted their methodology

In this thesis, we use the most common method which firstly was proposed by Narasimhan Jegadeesh and Sheridan Titman in 1993 to test the profitability of momentum (contrarian) strategies We form a new portfolio in each month t by going long 10% top stocks with highest performance and going short 10% top bottom stocks with lowest performance

This method consists of four following steps

The first step: At the ending of each month t during the selected sample analysis period from January 2005 to June 2011, we calculate the average monthly return of stock i on previous J months

#

"$#%&'(

)Where:

- ARit is average monthly return of stock i at month t;

- Riu is the log return of the stock i at the month u (month u: from month t-j+1 to month t);

- J is the formation period set to 3, 6, 9, or 12 months

The second step All stocks are ranked in descending order on the basis of their average

returns in the previous J months

The third step. Basing on rankings in the second step, we divided all stocks into ten equal portfolios The top decile is assigned the winner portfolio which contains the stocks with the highest past returns in the past J months, whereas the bottom decile is assigned loser portfolio which contains the stocks with the lowest past returns

The final step. We have to determine the profit difference between winner and loser The differene is equal the average monthly returns of winner portfolios ( *winner) minus the average monthly returns of loser portfolios ( *loser) in nexth K months

*winner-loser = *winner - *loser

The null hypothesis of the momentum (contrarian) strategies is that the differences between average returns on winner and loser portfolios R * winner-loser will be zero If the R * winner-loser are different from zero statistically significant, we can reject the null hypothesis Therefore, there is evidence of momentum (contrariance) phenomenon in the Vietnamese stock market, and the momentum (contrarian) strategies will generate significant abnormal profits In addition, we can also say that Vietnamese stock market is not weak form of Efficient Market Hypothesis

We use the above method for four different formations periods (J=3, 6, 9, 12 previous months) and four different holding periods (K=3, 6, 9, 12 next months) Consequently we have a total number of 16 strategies The process could be illustrated as follow:

Figure 3.1: Formation and Holding periods in two strategies

CHAPTER 4: EMPIRICAL RESULT

This chapter presents the profitability of contrarian (momentum) investment strategies which are applied to 424 stock tickers listed on Ho Chi Minh Stock Exchange and Ha Noi Stock Exchange over the period January-2005 to June-2011

4.1 Raw Data Processing

Before importing data to SPSS software to test null hypothesis of momentum (contrarian) phenomenon, we use the Analyzing Stock Price Data software to process raw data This program is composed by Nguyen Hiep Phat an Information Technology Engineer works

at Au Viet Securities Corporation

It imports return data from Excel file and gives an output of descending stocks which are ranking by average monthly returns in the previous J months

We have total of 16 strategies, and in each strategy we perform following steps:

In each month during the given period, all stocks are ranked ascending based on the average returns of the previous J months On the basis of that ranking, winner and loser portfolios are built The winner portfolios consist of the top decile which comprises the stocks with the highest performance in the previous J months, and the loser portfolios consist of the bottom decile which comprises the stocks with the lowest performance

Following image illustrates screen of the Analyzing Stock Price Data program after it import data from Excel file and ranking stocks in descending base on the 3M/3M strategy

in month 43 It imports monthly return data of 424 stocks from Excel file, after in every month this program give an output of descending return

Figure 4.1: The screen of the Analyzing Stock Price Data program after importing data

from Excel file

The following image illustrate stocks are ranked in descending order on the basis of their

average monthly returns in previous J months and their average monthly returns in next K

months

Figure 4.2: The screen of the Analyzing Stock Price Data after stocks are ranked in

descending order on the basis of their average monthly returns

Finally we use Stock Grouping to build the Winner and Loser portfolio and their average returns in every month Following image shows the Winner portfolio consisting of stocks with highest return and its average return (yellow line), and Loser portfolio consisting of stocks with lowest return and its average return (yellow line)

Figure 4.3: The screen of Stock Grouping program shows the Winner and Loser portfolios, and their average returns

Below table show the result the average monthly return of winner and loser portfolio in J=3/K=3 strategy after processing

Table 4.1: The average return of Winner and Loser Portfolios and their difference

in J=3/K=3 Strategy (Formation Period: 3 months; Holding Period: 3 months)

Winner minus Loser

Month

Average return of Winner Portfolio

Average return of Loser Portfolio

Winner minus Loser Apr-05 -0.0217 -0.0084 -0.0133 Apr-08 -0.0471 -0.0693 0.0222 May-05 0.0258 0.0271 -0.0013 May-08 0.0606 0.2665 -0.2059 Jun-05 0.0766 0.0278 0.0488 Jun-08 0.0227 0.1126 -0.0899 Jul-05 0.0557 0.0765 -0.0208 Jul-08 -0.0593 -0.0312 -0.0281 Aug-05 0.036 0.0442 -0.0082 Aug-08 -0.25 -0.1235 -0.1265 Sep-05 -0.0104 0.0326 -0.043 Sep-08 -0.1629 -0.1326 -0.0303 Oct-05 -0.0188 0.0239 -0.0427 Oct-08 -0.066 -0.0551 -0.0109 Nov-05 0.0148 0.0655 -0.0507 Nov-08 -0.0724 -0.1039 0.0315 Dec-05 0.049 0.156 -0.107 Dec-08 -0.0309 -0.0032 -0.0277 Jan-06 0.1358 0.209 -0.0732 Jan-09 -0.0121 0.001 -0.0131 Feb-06 0.1095 0.1575 -0.048 Feb-09 0.1294 0.2238 -0.0944 Mar-06 -0.0239 0.1456 -0.1695 Mar-09 0.1357 0.1156 0.0201 Apr-06 -0.1955 -0.0558 -0.1397 Apr-09 0.1357 0.0965 0.0392 May-06 -0.0771 0.001 -0.0781 May-09 0.0816 0.047 0.0346 Jun-06 -0.0496 0.0236 -0.0732 Jun-09 0.1229 0.0584 0.0645 Jul-06 -0.0053 0.0494 -0.0547 Jul-09 0.1198 0.1192 0.0006 Aug-06 0.0274 0.0323 -0.0049 Aug-09 -0.0051 0.0261 -0.0312 Sep-06 -0.0106 0.0294 -0.04 Sep-09 -0.0451 -0.0485 0.0034 Oct-06 0.1974 0.014 0.1834 Oct-09 -0.1131 -0.0823 -0.0308 Nov-06 0.2199 0.1464 0.0735 Nov-09 -0.0192 -0.0057 -0.0135

Jan-07 -0.0713 0.1508 -0.2221 Jan-10 0.0567 0.0923 -0.0356 Feb-07 -0.0707 0.0182 -0.0889 Feb-10 0.0319 0.044 -0.0121 Mar-07 -0.0693 -0.0507 -0.0186 Mar-10 0.0201 0.0224 -0.0023 Apr-07 -0.0156 -0.0246 0.009 Apr-10 -0.0412 -0.023 -0.0182 May-07 -0.0519 -0.1002 0.0483 May-10 -0.0788 -0.0407 -0.0381 Jun-07 0.0374 0.0562 -0.0188 Jun-10 -0.0812 -0.0621 -0.0191 Jul-07 0.1358 0.1767 -0.0409 Jul-10 -0.1375 -0.1115 -0.026 Aug-07 0.0823 0.2606 -0.1783 Aug-10 -0.06 -0.0349 -0.0251

We do the same approach for all 16 strategies (J/K), and have a following result in bellow table For each of the 16 strategies, a total of six summary statistics are shown: average monthly returns, average standard deviations for both the Winner portfolio and Loser portfolios Also the table 1 shows the average monthly returns Loser-Winner portfolio, and a test statistic for the significance of returns on the Loser-Winner portfolio

Table 4.2: Summary of the average monthly return of loser and winner portfolio; and their differences (profitability of contrarian strategies) for 16 strategies over the period from 01/2005 to 06/2011

(*) Significant at the 1% level

(**) Significant at the 5% level

(***) Significant at the 10% level

All of the average returns for Loser minus Winner portfolios for 16 strategies are positive and significant statistically at the level of at most 10% (the confidence level of at least 90%)

The two most profitable strategies are the 6M/3M and 3M/3M contrarian strategies with a Loser minus Winner portfolio that make a average monthly return of 2,829% and 2,306% respectively (equivalent to annually return of 33,95% and 27,67%) with significance level of 2% and 7% respectively

Table 4.3: Monthly and annually profitability of 16 contrarian strategies are ranking in descending and their significances

J/K contrarian

strategy

Average Monthly Return

Average Annually Return

Significant level (2-tailed)