D ETERMIN ATION MACROECON OMICS VARIAB LES AN D S TOCK RETU RN : A CAS E OF F IN AN CE S E CTOR AN D TRAD IN G & S ERVICE S E CTOR IN MALAYS IA

P a u lin e Ch ee

Ba ch elor of F in a n ce (H on ou r s) 2010

CHAPTER ONE INTRODUCTION

1.1

Introduction

Stock market is a place for listed companies to raise capital .Companies can use the capital for continuing operating activities and expand business. However, the investors are explained to get a positive return from dividend and capital gain in the stock market. Based on the history, the economic condition will influence stock market. For instances, Malaysia faced deflation during the Asian crisis in years 1997. It caused the KLCI index sharply reduced from 1207.43 to 470.43. It have been shown that the investors need to predict the stock prices based on the macro factors to get an abnormal return from stock market

There were a lot of researches to study the relationship between macroeconomics variables and stock returns. It is important to study the interaction of macroeconomics factor and stock return. Based on the study, the public can identify which factors can influence the stock market and use the knowledge to predict movement of stock price. According to Wongbangpo & Sharma (2002), the research can reveal the functions of stock market in identify the change in economic condition and also can predict the future performance of stock market. Besides, the study will be useful for the stock market participators. Clare & Priestley (1998) said that the study of the risk factor relationship of stock market will be useful for corporate manager to undertake cost of capital calculation. Moreover, the fund managers can use the information from the result of study to make an effective investment

decision and at the same times the investors can access the performance fund managers (Clare & Priestley, 1998). Lastly, the result of study can be used as guidelines for the government to implement policy. Butt et.al (2010) said that the finding of the study can be used to devise an effective economics and financial policy and improve the stock market condition in the country.

Basically, the previous study evidenced that stock return are determined by macroeconomics variables. Chen et.al (1986) found that industrial production, unanticipated inflation, expected inflation rate are significance to stock return. In the study of Azeez & Yonezawa (2006), the findings showed that money supply, inflation rate, exchange rate and industrial output have impact on stock return in Tokyo stock exchange. Moreover, Frimpong (2009) stated that inflation rate, interest rate and money supply have negative impact on Ghana stock market. The result from Gan et.al (2006) documented that money supply and inflation rate have significance impact on the New Zealand stock exchange. Besides, Eryigit (2009) examine the relationship between oil price and stock return. The study concluded that oil prices have positive significant in Oman, Saudi Arabia and Qatar. For a number of years, the investors and researchers have paid attention on emerging market especially in South Asia country. For examples, Wongbangpo & Sharma (2002) examined the interaction of macroeconomic factors on stock market for Indonesia, Malaysia, Philippine, Singapore and Thailand. The results showed that consumer price index has negative effect on stock market in the selected countries. Moreover, Maysami et.al (2004) found that inflation rate and money supply have positive relation on stock return in Singapore stock market. Rahman et.al (2009) document that the money supply has negative effect on Malaysia stock market but industrial product has positive relationship with stock return.

The emerging market have distinguish feature from developed market. According to Harvey (1995), the emerging market provides high expected return with more predictability power compare with developed country. The specialty feature of emerging market provides an investment opportunity to investors. In fact, Malaysia stock market is an emerging stock market. Exception for Asian crisis and Mortgage crisis in year 1997 and year 2008, Malaysia experienced a growth in GDP rates in past few years. During years 2002 to years 2010, the GDP rate of Malaysia increase from 5.39% to 6.18%. Therefore, it is interested to examine the interaction of macro factor on Malaysia stock market since the stock market provides high potential of growth.

The general objective is to examine the interaction between macroeconomics factor and sector stock return. There are three specific objectives which are to examine the interaction between stock return, industrial production, inflation rate, oil price and money supply for finance sector, to examine the relationship between industrial production, inflation rate, oil price and money supply for trading and service sector, and lastly to find out whether the interaction of stock return, industrial production, inflation rate, oil price and money supply in different sector.

There are four macroeconomics variables which consist of industrial production, inflation rate , oil price and money supply are obtained in the study. The study examines the interaction of stock return and macro variables by using sector approach. The study focuses on finance sector and trading & service sectors in Malaysia. The reason of selected trading &

service sector and finance sector are because both sector contributed the high GDP growth in 2009 and 2010. The data for the variables obtained for the period of 13 year. The period from January 1998 to December 2010 has been selected in the study. Data was collected from Thompson Reuters Datastream at Unimas and Bank Negara Malaysia.

In term of methodology, the study employed unit root tests consist of Augmented Dickey Fuller test, Philip Perron test and Kwiatkowski-Phillip Schmidt-Shin test to analyze the stationary of the times series of data. The finding of the unit root tests show that all the times series data are stationary and integrated order I(1). Besides that, Johansen Cointegration test will be used for investigate the relationship between selected variables and stock return in long term. The result indicates that there are one cointegration factor in the trading & service sector. However, the result show that he finance sector does not have cointegration factor in the long run. Lastly, the VECM Granger Causality test and Granger causality test will be conducted to examine the relationship between the macro variables and stock return in short run. The finding shows that oil price and inflation rate have significance impact on stock return in trading & service sector. Besides, the result also concludes that there are short run relationship between money supply and stock return in finance sector.

The study contributes to the literature in two ways. Firstly, the study provides a sectoral measurement of stock return caused by macroeconomics factors. The study investigates the interaction between macroeconomics factor and stock return in finance sector and trading & service sector. Secondly, the finding of the study will be meaningful for investors and policy makers. The study provides useful information regarding finance sector and trading &service sector. Through the information, the investors have deeper knowledge

about the movement the two sectors. Therefore, the investors can take the advantages to gain abnormal profit by make an efficient investment portfolio. The reminder of the study is organized as followings. The chapter two explains the theoretical framework and the previous studies. The data employed and methodologies are described in chapter three. The chapter four will discuss the finding. Finally, the chapter five will provide that conclusion of the study.

1.2

Background of Study

Malaysia

Malaysia is a federal constitutional monarchy in Southeast Asia. It consists of thirteen states and three federal territories. Malaysia is divided into two regions which consist of peninsular Malaysia and Malaysia Borneo.

The Figure 1.1 shows that the GDP has consistent growth from 1961 to 1972. In 1971 and 1972, the increase oil price has contributed the GDP growth from 5.7% in 1971 to 11.7% in 1972. Malaysia economy happened recession in 1975. The GDP have showed a slow growth 0.8% in 1975. During 1977 to 1984, the GDP has consistent growth. In 1985, the downturn of electronic sector has negative impact on Malaysia economics (Treasury Malaysia, Ministry of Finance, 2010). The GDP has reduced from 7.76 % in 1984 to -1.12% in 1985. Malaysia kept high interest rate to attract foreign investment. The higher interest rate attracts a huge amount of foreign capital inflow (Treasury Malaysia, Ministry of Finance, 2010). Malaysia experienced high growth from 1994 to 1996. The GDP has increased from

9.21% in 1994 to 10% in 1996. However, the GDP has reduced to 7.3% in 1997. Malaysia faced the Asian economic crisis in 1997. The economics Malaysia experienced low growth which GDP become -7.3%. The GDP of Malaysia has recovered to 6.13% in 1999. The exchange control and pegging of the ringgit at RM 1 = US$ 0.2632 has brought confidence in the market (Treasury Malaysia, Ministry of Finance, 2010). Besides that, the positive growth in GDP was contributed by recovery in the manufacturing and agricultural sector and the improvement of oil yield (Treasury Malaysia, Ministry of Finance, 2010).

In 2001, the GDP growth became 0.5%. The slow performances in United States and Japanese economy have slowdown the performance of Malaysia economy (Treasury Malaysia, Ministry of Finance, 2010). During 2002 to 2007, Malaysia has positive growth in GDP. The GDP has increase from 5.39 % to 6.18%. The constant growth of Malaysia economics is contributed by the increasing demand for housing, motor vehicles, and electronics (Treasury Malaysia, Ministry of Finance, 2010). In 2008, the GDP has reduced to 4.6%. The economy of Malaysia faces a negative growth -1.7% in 2009. The slow economic performance is affected by the slowdown of United States economy which arising by the subprime mortgage crisis (Treasury Malaysia, Ministry of Finance, 2010). Figure 1.1: Gross Domestic Product Growth (%) in Malaysia from Year 1961 to Year 2009

Notes: World Bank (2010)

KLCI Stock Exchange

The first formal securities business organization in Malaysia was established in 1930 (Bursa Malaysia Berhad, 2010). It was namely as Singapore Stockbrokers Association. In 1937, the Singapore Stockbrokers Assiociation was re-registered as the Malayan Stockbrokerss Association and the public trading of shares are not traded (Bursa Malaysia Berhad, 2010).

The Malayan Stock Exchange was introduced in 1960 (Bursa Malaysia Berhad, 2010). The public trading of shares can commenced on Malayan Stock Exchange. In the years 1961, the board system was introduced. There are trading room in Singapore and Kuala Lumpur which linked by telephone links (Bursa Malaysia Berhad, 2010).

The stock exchange of Malaysia was incorporation in 1964(Bursa Malaysia Berhad, 2010). In 1965, Singapore was secession from Malaysia. The stock exchange of Malaysia became stock exchange of Malaysia and Singapore (Bursa Malaysia Berhad, 2010). Due to the currency interchangeability between Malaysia and Singapore, the Kuala Lumpur Stock Exchange Berhad was official split from stock exchange of Malaysia and Singapore in 1973(Bursa Malaysia Berhad, 2010). The Kuala Lumpur stock exchange was established on December 14, 1976(Bursa Malaysia Berhad, 2010).

On April 14, 2004, the Kuala Lumpur Stock Exchange changed their name to Bursa Malaysia Berhad(Bursa Malaysia Berhad, 2010). Bursa Malaysia Berhad continually offering full- integrated exchange and provide complete exchange related services. Bursa Malaysia Berhad set up wholly-owned subsidiary. On 18 March 2005, Bursa Malaysia Berhad has set a Bursa Malaysia Securities Berhad (Bursa Malaysia Berhad, 2010). The Bursa Malaysia Securities Berhad provides, operate and maintain securities exchange (Bursa Malaysia Berhad, 2010). Besides Bursa Malaysia securities Bhd, Bursa Malaysia also set up several businesses. There are subsidiary businesses including Bursa Malaysia Derivatives, Labuan International financial exchange, Bursa Malaysia Bonds Sdn Bhd, Bursa Malaysia Securities Clearing Bhd, Bursa Malaysia Derivatives Clearing Bhd, Bursa Malaysia Depository Sdn Bhd, Bursa Malaysia Depository Nominees Sdn Bhd and Bursa Malaysia Information Sdn Bhd(Bursa Malaysia Berhad, 2010).The Bursa Malaysia provides under 1000 listed companies and those companies can be listed on Bursa Malaysia Securities Berhad main market or ACE market.

Bursa Malaysia introduced Kuala Lumpur Composite (KLCI) in 1986. The index is known as FTSE Bursa Malaysian Malaysia KLCI. The Figure 1.2 shows the KLCI index closing price from December 1994 to December 2009. In 1997, the KLCI index decline from 1207.43 to 470.43. The dramatically declined is caused by Asia Economic crisis. The KLCL index increase from 1096.24 to 1445.03 in December 2007. The high growth of KLCL index is contributed by the high growth of U.S economy. However, the subprime mortgage crisis leads to reduce the KLCL index to 876.75 in December 2008.

Figure 1.2: KLCI Index Closing Price

Notes: Yahoo Finance (2010)

As at September 2010, the Bursa Malaysia Index Series were classified into nine sectoral major indices, namely, construction, consumer product, finance, industrial product, mining, plantation, property, technology, and trading & service.

The sectoral finance, mining, plantation and property indices were established in 1970s. As at 30 September 2010, the sectoral indices consist of 38 stocks. For the mining

sectoral indices, it consists of 1 stock which known as Kuchai Development Bhd and sectoral plantation included 43 stocks. There are 87 stocks in sectoral of property. In 1992s, Bursa Malaysia introduced its new 4-sector classification of listed companies which included construction, consumer product, industrial product and trading services. The construction sector has 50 stocks at September 2010 and sector consumer product included 139 stocks. companies. The sector trading service included 179 stocks of listed

The sector industrial product is the largest sector indices in Bursa saham

Malaysia. It consists of 263 stocks of listed companies. The Bursa Malaysia introduced sector technology in 1999s. The sector technology included 29 stocks in Bursa Malaysia.

1.3

Motivation of Study

There are two factors that motivate a study. First, unlike the previous studies that focuses on identified effect on stock return from developed countries, the study focuses attention to find out the interaction between stock return and macroeconomic factors in Malaysia stock market. As an emerging market, Malaysia stock market has provided an investment opportunity to public. The findings of study will provide additional information regarding Malaysia stock market. It will increase the understanding and knowledge among public towards Malaysia stock market. Thus, it also helps to attract more foreign investor to invest in Malaysia. Since previous studies on Malaysia have been conducted by Clare & Priestley (1998), Wongpo & Sharma (2002) and Rahman et al (2009), however there were less studies examine stock return and macroeconomics factors by sector approach. Therefore,

the study will focus on two major sectors which is finance sector and trading & service sector in Malaysia stock market. The two sectors have distinguished features and growth. Therefore, the interaction of stock return and macro variables might be different in two sectors. Based on the findings of study, the investors can identify which macro factors can be used to predict stock return in different sectors. Thus, the investor can diversify their investment portfolio in different sectors based on their risk level.

1.4

Problems Statement

The study of the relationship of stock return and macroeconomics always been studied among the academics. In general, the macroeconomics variables always are considered as factors that affect the stock return. Therefore, it implied that the investor will change their investment planning based on the changes of the macroeconomics condition. Based on those reason, it motivated many academics to examine the effect of macroeconomics on stock return. There are a number of study investigates the interaction between macroeconomics factor and stock return in several countries such as London, United States, Brazil, India, China, Singapore, Thailand, and Malaysia. The founding from the previous study are still under the consideration of mix concluding remark among the macroeconomics. Thus, it is interested to find out whether there is any relationship of stock return, industrial production, inflation rate, oil price and money supply in Malaysia. From the previous studied such as Tursey et.al (2009) and Gunsel & Cukur (2007), they stated that the macroeconomics factor may affect one industry positively or negatively. In Malaysia, all sectors have different structure and characteristic. The Performance of sector might be different based on the change of economics condition and business environment. According to Ministry of finance (2010), the economics report in 2010/2011 show that the

steady performance Islamic financing and new business contribution for takaful have contributed high GDP rate in finance sector. In contrast, the trading and service sector expanded GDP due to robust trade and consumption activities. The question to be addressed is: Are there any different relationship between macroeconomics factor and stock return for finance sector and trading & service sector.

1.5

Objectives of Study

1.5.1 General Objective The main objective of the study is to examine the interaction between industrial production, inflation rate, oil price money supply and stock return for finance sector and trading & service sector in Malaysia stock market.

1.5.2 Specific Objective. The specific objectives include: i. To examine the relationship between stock return, industrial production, inflation rate, oil price and money supply for finance sector. ii. To examine the relationship between stock return, industrial production, inflation rate, oil price and money supply for service sector. iii. To find out whether the interaction of stock returns, industrial production, inflation rate, oil price and money supply in different sector.

1.6

Significance of Study

The study investigates the relationship between stock return and macroeconomics factors for finance sector and trading service sector. In the previous study, the researches mostly study about relationship between macroeconomics factor and stock return by using the country approach. It is less attention to studies about the effect of Malaysia stock return by using sector approach. Thus, the study can be considered as an important study to investigate the factor that effect on sectors in Malaysia stock market.

Besides, the study will be beneficial for various professional and market participation that interested to invest in Malaysia stock exchange. Based on the study, the investors can identify the macroeconomics variables that can influence stock return. The understanding of relationship of macroeconomics factor and stock return is important for the investors. An accurate estimation of relationship of stock return and economics risk will enable for investor to make efficient investment decision. Besides that, the result also can help regulatory bodies to devise a better financial policy to improve stock market condition in Malaysia.

1.7

Organization of Study The study emphasize on the relationship between the relationship between stock

return, industrial production, inflation rate, oil price and money supply in the selected industry in Malaysia stock exchange from January 1998 to December 2010.

The study is organized as following: The chapter two reviews the related previous studies have been done by researches. The chapter two divided into four sections which included theoretical framework, empirical testing procedures, empirical evidence and summary of previous studies. The chapter three discuss about the data and method that been

use in the study. Chapter four is used to analysis and interprets the empirical result. Finally, the chapter five concludes the overall finding and provides policy recommendation.

CHAPTER TWO LITERATURE REVIEW

2.1

Introduction

In generally, the stock return believes to be effected by fundamental macroeconomics variables. There were several researches study the interaction between stock return and economics factors in many countries such as Japan, Malaysia, and New Zealand.

Chen et al. (1986) developed arbitraged pricing theory to examine the effect of a several macroeconomics variances on US stock return. The selected macroeconomics various included industrial production, change in expected inflation, expected inflation, risk premium, term structure, consumption and oil price. Based on their study, they found that industrial production, risk premium, term structure, unanticipated inflation, expected inflation is significant to the stock return. Besides that, Chen et al. (1986) also indicated that consumption, oil price and market index are not significant to stock return. The study explained that the stock returns are exposed to systematic news. The applicability arbitraged pricing theory have been tested by several researchers such as Tursoy et al. (2008), Gunsel & Cukur (2007) and Azeez & Yonezawa (2006). The empirical results regarding to relationship of macroeconomics variables and stock return have been discussed in this chapter.

Chapter two divided into six sections. Section 2.2 will be discusses the theoretical framework of study. The theoretical model will be discussed in Section 2.3. Section 2.4 will be discussed about the empirical testing method used in the previous studies. The empirical result will be explained in section 2.5. A summary of our finding is provided in Section 2.6.

2.2

Theoretical Framework

Based on the previous study, the relationship between industrial production, inflation rate, money supply, oil price and stock return can be identified as following;

2.2.1 Inflation Rate

The prior study showed that inflation rate has negative effect on stock return. The higher inflation rate will increase cost of firm. It will lead to decrease future cash flow and revenue of firm. The stock return will decrease due to the reducing of future cash flow. The negative relations of inflation rate and stock return have been proven by Frimpong (2009) and Gan et al. (2006). Frimpong examine the relationship of macroeconomics factor and stock return in Ghana stock exchange by using Johansen cointegretion test and vector errorcorrection model. The result showed that inflation rate has negative effect on stock return. The finding of Gan et al. (2006) is similar with Frimpong (2009). The study of Gan et al. found that inflation rate has negative relationship with stock return in New Zealand stock index.

2.2.2 Money Supply

The literature study concluded that the money supply can be negative effect or positive effect on stock return. When the money policy is not credible, the money supply will bring negative effect to the stock return through the uncertainty of inflation. The study of Rahman et al. (2009) concluded that money supply has negative effect on stock return of Malaysia. Gan et al. (2006) applied innovation accounting analysis to examine the effect of money supply on

stock return. The result showed that money supply has negative impact on stock return in New Zealand stock index. However, there were previous studies found a positive relationship between money supply and stock return. The finding of Wonbangpo & Sharma (2001) & Maysami et al. (2004) showed that the money supply has positive relationship with stock return in Singapore, Malaysia and Thailand. Money supply also affect on stock return. The increasing money supply can boost up the economics and increase the real activity. Therefore, the stock return will also increase.

2.2.3 Industrial Production

The previous study stated that there are positive relationship between industrial production and stock return. The industrial production affects the stock return through future cash flow. The increasing of industrial production will increase the revenue and profit of firm and thus increase cash flow of firm. At the same times, the increasing cash flow will lead to increase stock return. The positive relationship between industrial production and stock return has been proven by Rahman et al. (2009). Rahman et al. examined effect of economics factor in Malaysian market. Their study showed that industrial production has positive relationship with stock return. The study apply Ganger causality test in VECM and the test showed that the industrial production have long run effect on Malaysias stock return. Besides Rahman et al. (2009), Aburgi (2006) also study the stock return in Latin American market by applied Vector Autoregression Model. Based on their study, they found that there was a positive relationship between industrial production and stock return in Brazil and Chile.

2.2.4 Oil Price

The prior study identified that the oil price have positive effect on stock return. The oil price may influence stock return through future cash flow. The increasing oil price can increase government and corporate revenue and thus can increase future cash flow and stock return. Arouri et al. (2010) observed that oil price is positive effect in Oman, Qatar, and Saudi Arabia and UEA stock market. Erygit (2009) identified stock return by selected 16 sectors in Istanbul stock exchange from 2000 to 2008. The study found that the oil price has positive effect on wood, paper & printing, insurance and electricity sub-sector indices.

2.3

Theoretical Model

Arbitrage Pricing Theory was developed by Ross (1976). Ross (1976) stated that if the prices do not offer arbitrage opportunity over static portfolio assets, it mean that return on assets are related to factor loading. According to Ross (1976), there are three major assumptions in APT. Firstly, capital market is perfectively competitive. Second, investors prefer more wealth with certainty. Lastly, the assets return can be express as linear function of K risk factor. Ingersoll (1984) argue that APT do not give any cue for identify the significance factor. In the past decades, a number of researchers have proposal new methodology for tested APT. The macroeconomic-Based risk factor model was developed by Chen et.al (1986) .The model was include the essence of APT which to estimate a relationship of specification form. Chen et al. tested the relationship between industrial production, change in expected inflation, expected inflation, risk premium, term structure, consumption and oil price. The model express as following: (2.1)

where R is the real return on the stock and a is the constant term and b is the reaction coefficient measure the change in return for a change of macroeconomic factor . The set of macroeconomics various include MP: industrial production DEI: change in Expected inflation UI: Unanticipated inflation UPR: Risk premium UTS: Term structure e: idiosyncratic error term The equation (2.1) implies that change of industrial production, expected inflation, unanticipated inflation, risk premium and term structure can affect the exchange rate. Chen et.al (1986) hypothesized that industrial production has positive effect on stock return. The study assume the risk premium have positive relationship with stock return. It is because the individual want to invest against increasing risk premium. The inflation rate expected to be negative sign on stock return.

Frimpong (2009) extended the study of Chen et al. (1986). The model be written as following (2.2) represent the

Where

represent as the error term,

defined as constant term,

long run parameter operator,

is the natural logarithm of exchange rate, is the natural logarithm of interest defined as GSE All-share

is the natural logarithm of Inflation rate, rate, index.

is the natural logarithm of money supply and

The equation (2.2) implies that logarithm of exchange rate, logarithm of inflation rate, logarithm of interest rate and logarithms of money supply have relationship with logarithm of stock return in GSE index. When the logarithm of exchange rate, logarithm of inflation rate,

logarithm of interest rate and logarithms of money supply increase, the logarithm of stock return in GSE index will also increase.

2.4

Empirical Testing Procedures

Previously, the Unit Root tests, Johansen Cointegration test, ARDL Cointegration Test, VECM Granger Causality test, Granger Causality test have been conducted in the prior studies. This sector will discuss those test conducted by the previous studies. The unit root was conducted to examine the stationary of each series. The unit root test will be further discussed in section 2.4.1. For the Johansen Cointegration test and ARDL Cointegration test were conducted to examine long run relationship between the series. The Johansen Cointegration test and ARDL Cointegration test will be discuss in section 2.4.2 and 2.4.3. The VECM Granger Causality test and Granger Cointegration was implemented in previous study to examine the short run relationship between the variables and it will be discuss in section 2.4.4 and 2.4.5.

2.4.1 Unit Root Test

Three types of unit root tests were conducted to identify the stationary of times series data. It consists of Dickey-Fuller Unit-Root Test, Phillip-Perron Test and KwiatkowskiPhillip-Schmidt-Shin (KPSS).

Augmented Dickey-Fuller Unit-Root Test

The Augmented Dickey- Fuller Unit-Root Test have been applied in the study of Butt et.al (2010), Mahmood & Dinniah (2009), Maysami et al. (2004) , Rahman et al. (2009) , Aktham (2004), Gay (2008) and Bilson et al. (2001). The Dickey-Fuller unit root test is used for identify whether the variables is stationary or non-stationary. According to Mishra (2010), the equation can be written by (2.3) (2.4) (2.5) Where coefficients, is differencing of variables of return, is the times series trend, is the intercept, and denoted as is the error

is the number of lagged term and

term. Model (2.3) is a pure random in lag term, model (2.4) include intercept and lastly model (2.5) include trend and intercept.

The hypotheses for the ADF unit roots test are:

The null hypothesis implies that the existence of unit root. The variables are considered non-stationary. In contrast, the alternatives hypothesis means that does not contain unit root test which considered that the series is stationary.

Mahmood & Dinniah (2009) tested ADF test and the result showed that it is fail to reject null hypothesis for all selected countries with exception of industrial output for Japan and inflation for Thailand.

Phillip-Perron Test

The Phillips- Perron test was introduced by Phillip & Perron (1988). The PP test was used to modify the ADF test. The result of PP is similar with ADF test. According to Chen et.al (2002), the distinguish between the PP test and ADF test are the PP test provide wide range of serial correlation and time-dependent heteroskedasticity.

Chen et.al (2002) stated that the regression for PP test can be expressed as (2.6) Where is the times series represented as defined as the innovation term, is represent

the number of observation Similarity with ADF test, the hypothesis for PP test is

. The null hypothesis implies that existence of a unit root. Therefore, the series are considered non-stationary. The alternative hypothesis implies that the series do not contain unit root and the variables is stationary.

The PP test have been applied in the study of Frimpong (2009), Rahman et al. (2009),Aktham (2004) and Bilson et al. (2001). The result from Frimpong (2009) showed that the log of exchange rate, money supply, inflation rate and interest rate are stationary at the level I (1).

Kwiatkowski-Phillip-Schmidt-Shin (KPSS)

The Kwiatkowski-Phillip-Schmidt-Shin (KPSS) was developed by Kwiatkowski et.al (1992). Choudhry (2001) stated that the KPSS test is valid for small sample and able to distinguish short memory from stationary long memory. According to Henry & Shields (2004), the regression for KPSS can be express as (2.7) where represent as partial sum of residual , defines as error variance , is the lag

parameter and =0

is the number of observation. The hypotheses for KPSS test are

The null hypothesis means that non-existence of unit root and variables are stationary. The alternative hypothesis indicate that exist unit root and the series are non-stationary.

The KPSS test has applied in several studies such as Rahman et al. (2009), Aktham (2004) and Choudhry (2001). In the result of Rahman (2009), it show that KLCI series, Money supply, Industrial production, exchange rate, Treasury bill are stationary at first difference.

2.4.2 Johansen Cointegration Test Gan et al. (2006) applied the Johansen Cointegration test to investigate long run relationship of variables. The Johansen Cointegration test can be explain by (2.8) where for i = 1,2,K-1; is an identity matrix where the denoted as short term adjustment parameters and comprise long-term

equilibrium relationship X variables.

decomposed into the product of two n by r matrix

and . Therefore

is a matrix contain r cointegration vectors and denoted as the

speed of adjustment parameter.

Two likelihood test used for examine the number of cointegration vector. The likelihood test consists of trace test and maximum Eigenvalue test.

Trace Test

The equation can be express as following (2.9) Where T is the number of observation, N is number of variables, and estimated eigenvalue. is the largest

According to Gan et al. (2006), the hypothesis of trace can be explained as following; R=0 R 0

The null hypothesis indicates that there is no cointegration. In contrast, the alternatives hypothesis implies that there is one or more cointegration vector.

In the study of Gan et al. (2006), the result of Johansen Cointegration test found that the variables are cointegrated with r = 6 at 5 % significance level. It implied that there are 6 cointegression vectors. The Johansen cointegration are widely used in previous studied such as Wongbangpo, & Sharma (2002) ,Frimpong (2009), Azeez & Yonezawa (2006), Mahmood & Dinniah (2009), Rahman et al. (2009)

Max Eigenvalue Test

The Eigenvalue Test can define as following; (2.10) where T is the number of observation and is the largest estimated eigenvalue .

Gan et al. (2006) stated the hypotheses are cointegrating vector cointegrating vector The null hypothesis indicates that there is no cointegration. In contrast, the alternative hypothesis implies that there is one or more cointegration vector. In the study of Gan et al. (20006), the Max Eigenvalue showed that at least four cointegration vector in model.

2.4.3 ARDL Cointegration Test

The bound testing approach of co-integration was advanced by Pesaran et.al (2001). The ARDL test has more econometric advantage in comparing with other single cointegration procedure. Halicioglu (2004) stated that the ARDL test have several advantage. First, it can avoided endogeneity problems. Second, long and short-run parameter of model is estimated. Thirdly, allow that the order of integration of variables may not be same which can be mixture of I (0) and I (1).

Based on Atif et.al (2010), an ARDL was formulated as following, (2.11)

Where

is vector of both

and

and

defined as economic growth. xt is represent a set

of explanatory variables.

In the case of Atif et.al (2010), after develop vector error correction model, it can be formulated as (2.12) where is the first-difference operator.

In order to unrestricted intercept and no trend, the unrestricted error correction model (VECM) will formulated as

(2.13)

Where

is the first-difference operator. The GDP defined as Gross Domestic product, TOP

defined as trade openness and M2 represent as broad money supply.

The ARDL approach will use the Wald test (F-statistic) to determine the long run relationship between selected variables. According to Atif et.al (2010), the null and alternative hypotheses can be express as

where

defined as no long run relationship.

defined as long run relationship

Halicioglu (2004) stated that one set of variables will be assume as I (0) and other set will considers as I(1). If the computed F-statistic is larger than upper critical value, therefore the null hypothesis is rejected. Thus, it can be conclude that there is long run relationship

between the variables. However, if the computed F-statistic smaller than lower critical value, therefore should not reject Ho. It can conclude that no long run relationship between variables. In the other situation whereby the computed F-statistic in between lower and upper bound value, it can be conclude as inconclusive.

2.4.4 Vector Error-correction Model (VECM) Granger Causality test

Wongbangpo & Sharma (2002) employed granger causality test based on vector error correction model (VECM) to examine the causality relation between variables. The equation can be express as following;

(2.14) where length, is the error correction term and are parameters , K is the lag

is the stationary random process with mean zero and constant variance. The null . It indicated that the stock return do not

hypothesis is Granger-cause selected variables.

The result of test in the study of Wongbangpo & Sharma (2002) found that stock market granger- cause to gross national product and consumer price index in Indonesia, Malaysia, Philippines, Singapore and Thailand. The result also showed that the stock market granger-cause to money supply and interest rate in Indonesia, Thailand and Malaysia and there are granger-causality to exchange rate in Philippine and Singapore.

2.4.5 Granger-Causality Test

Gan et al. (2006) applied Granger-causality test to investigate the relationship of leadlag between NZSE return and selected macroeconomics. The study stated that if variables is nonstationary and cointegrated , the test will be run by the following equation: (2.15) (2.16) where and and is the first different of times series variables when the series is denoted as stationary random processes to capture relevant and . and are measuring the error

nonstationary.

information that not include in lagged value of correction mechanism that drive the and

to long run equilibrium relationship.

However, if the variable is non-stationary and is not co-integrated, the Grangercausality test can be express by (2.17) (2.18) The null hypothesis for the equation (2.7) and (2.9) is : (the lagged term do not arise in regression)

The null hypothesis for the equation (2.8) and (2.10) is : ( the lagged term do not arise in regression)

The result of Grange- causality test from Gan et al. (2006) showed that exchange rate, short term interest rate, domestic retail oil price, and gross domestic product can be used for determine stock return. The result show that there was 5% of exchange Rate and short term interest rate cause with NZSE20 stock return. The domestic retail oil price can explained stock return in 1%. The Gross domestic product was co-integrated with stock return in 10%. The

Granger-causality test has been applied in Wongbangpo, & Sharma (2002) and Rahman et al. (2009).

2.5

Empirical Evidence

Chen et al. (1986) determined the macroeconomics factor and stock market from January 1953 to November 1983. The study employed several variable which included Industrial production, Unexpected inflation, Expected inflation, Term structure, Oil price, Consumption, Risk premium. The study found that Industrial production, risk premium, term structure, unanticipated inflation, expected inflation is significant to stock return. However, the study found that consumption, oil price and market index are not significance to stock return. The no relation between oil price and stock return has been supported by Aktham (2004) and Gay (2008).

Gay (2008) have examined the effect of exchange rate and oil price on stock market return for Brazil, Russia, India, and china from the period of 1999 to 2006. The result show that there are not significant effect of exchange rate and oil price on stock return for Brazil, Russia, India, and China. Aktham (2004) found a similar result with Gay (2008). The study showed that crude oil price do not effect on stock return of emerging countries. The selected emerging countries included Argentina, Brazil, Chile, China, Czech Republic, Egypt, Greece, India, Indonesia, Jordan, Korea, Malaysia, Mexico, Morocco, Hungary, Pakistan, Philippines, Poland, South Africa, Taiwan, Thailand, and Turkey.

The results of Chen et al. (1986), Aktham (2004) and Gay (2008) have been argued by Arouri et al. (2010) and Eryigit (2009). They concluded that oil price may significance and no significance on stock return. The study of Arouri et al. (2010) employed oil price to identify the stock return in GCC countries which included Qatar, Oman, Saudi, Arabia and UAE. The finding showed that oil price is positive significance in Oman, Qatar, Saudi Arabia and UEA. However, their study also found that Oil price are not effect in Bahrain and Kuwait stock market.

The relationship between oil price and stock return also has examined by Eryigit (2009). The study identified the effect of oil price on stock return by employed 16 selected sectors in Istanbul stock exchange from 2000 to 2008. The result conclude that the oil price are significance in electricity, wholesale and retail trade, insurance, holding, investment, wood, paper, printing, basic metal, metal products, machinery and non-metal and mineral product. In contrast, it found that there are no significance on transportation, banks, leasing, factoring, and real estate investment trust sector.

There are few studies of Istanbul Stock Exchange have been done by Tursoy et al. (2008) and Erdogan & Ozlale (2003). The study of Erdogan & Ozlale (2003) showed that before 1994 (financial crisis) the foreign exchange rate have positive effect on stock return. However, the relationship of foreign exchange rate and exchange rate become negative after crisis. Secondly, the industrial production are positive significant to stock return. The relationship have excluded the period from 1994 financial crisis and end with the beginning of 1997 Asian crisis. Thirdly, the interest rates influence the stock return in positive way until 1994 financial crisis.

Tursoy et al. (2008) applied Ordinary least square (OLS) technique to Examined 175 stock which are classified in 11 industry sector and listed in Istanbul Stock Exchange from January 2001 to September 2005. The study states that the macroeconomic factor may affect one industry positively or negatively.

In the study of Chen et al. (2005), the result showed Money supply and unemployment rate have significant to the hotel stock return in Taiwan. Gunsel & Cukur (2007) examined the 10 industry sector and listed in London Stock Exchange for period 1980 to 1993. The result found that the money supply has positive effect on the return of Building material & Merchants, Food, Beverage and Tobacco. In contrast, there are negative effect on the return of household goods and textiles. The inflation rate only effect on return of food, beverage and tobacco and the interest rate have positive relationship with construction, food, beverage & tobacco, oil exploration & production and electronic & electrical equipment industry.

Bilson et al. (2001) identify the macroeconomics factor that effect on stock return for 20 emerging countries from 1985 to 1997. The study employed Dickey Fuller, Phillip-Perron test and correlation test. The result showed that the exchange rate is significance with the return for twelve markets. However, the money supply, real activity and good prices are not strongly effect on the return on emerging market.

The no relations between money supply and stock return have been supported by Abugri (2008). The study of Abugri examined the interaction between macroeconomics factor and stock return for four Latin American countries. The result showed that the impact of money supply and industrial production toward stock market in Argentina, Brazil, Chile and Mexico are weak. The study further explained that the interest rate have significance effect on

stock market for Brazil, Mexico, and Argentina. The stock return responds significance to exchange rate in the stock market in Brazil, Mexico.

Gan et al. (2006) examined relationship between the macroeconomics factor and stock market in New Zealand stock exchange form January 1990 to January 2003. The study have selected seven macroeconomics variables which included share price Index, inflation rate , exchange rate, gross domestic product, long term interest rate, short term interest Rate, domestic retail oil price. They employed Johansen Maximum likelihood, Granger-causality test and innovation accounting analysis in the study. The result of Johansen Maximum likelihood and Granger-causality test show that New Zealand Stock Index 40 which is not a leading indicator for the changes in macroeconomics variances because New Zealand stock market is small if compare to other developed countries. Gan et al. (2006) also concluded that the exchange rate, inflation rate, long term interest rate, gross domestic product and money supply have positive effect on New Zealand Stock Index. However, the study showed that the money supplies have negative effect on stock index.

Furthermore, in a study of ASEAN countries such as Wongbangpo & Sharma (2002) conclude that consumer price index effect negatively on stock return in all stock prices in all ASEAN countries. Money supply has positive effect on stock price in Malaysia, Singapore and Thailand. However, the negative effect of money supply on stock price can be observed in Indonesia and Philippine. The interest rates have negatively influence stock market in Philippines, Singapore, and Thailand. Similarly with Wongbangpo & Sharma (2002), the result of Maysami et al. (2004) showed that money supply has positive relationship with stock return in Singapore stock index.

Mahmood & Dinniah (2009) investigate in Asian-Pacific countries from January 1993 to December 2002.The selected countries included Indonesia, Malaysia, Philippines, Singapore and Thailand. The Engle-Granger test show that Thailand stock return correlate with foreign exchange and inflation while Japanese stock return cointegrated with inflation rate and industrial output. The Korea stock return cointegrated with foreign exchange and Hong Kong stock return cointegrated with inflation. The study also concluded that Thailand and Hong Kong have relationship between stock market and macroeconomics variables in the short run.

Azeez & Yonezawa (2006) concluded a similarly result with Mahmood & Dinniah (2009). Azeez & Yonezawa (2006) tested 28 industries in Tokyo Stock Exchange from January 1973 to December 1998. The finding of study showed that inflation rate and industrial output are significance to Japanese stock market. Besides that, the study found that money supply and exchange rate have significance effect on stock return.

In Malaysia, Rahman et al. (2009) examine effect of interest rate, reserve, industrial production, exchange rate, and money supply by apply several test which consist of Augmented Dickey Fuller, Philip Perron test, Kwiatkowski- Philip-Schmidt-Shin, JohansenJusellus test and Granger causality test. The result shows that industrial production and reserve have positive relationship with stock return. However, the money supply and exchange rate, interest rate has negative effect on stock return.

A similarly study in Malaysia have been examine by Clare & Priestley (1998). The study examined 70 stocks from Malaysia stock exchange from January 1986 to August 1994. Different with the result of Rahman et al. (2009), the study found that industrial production is

not effect on the stock return. It is only the unexpected changes in risk free rate, unexpected changes in term structure of interest rates, unexpected inflation, and changes in expected inflation are significant on stock return

2.6

Concluding Remark

Over the past few decades, the macroeconomics and stock return have been investigated by many researches. The studies have been done in European countries and Asian countries.

As summarize, the study of Gan et.al (2006) and Frimpong (2009) showed that inflation rate have negative impact on stock return. However, Chen et al (2005) and Butt et.al (2010) indicated there was positive relationship between inflation rate and stock return.

For the series of money supply, the finding from Gan et.al (2006) and Rahman et.al (2009) concluded that there was negative significance to stock return. In contrast, Abugri (2008), Butt et.al (2010) and Bilson et.al (2001) concluded that money supply have positive impact on stock return.

Erdogan & Ozlale (2003) documented that industrial production have positive relationship with stock return but the study from Rahman et.al (2009) concluded that there was negative relationship between industrial production and stock return. Based on the study from Gan et.al (2006) and Aroubi et.al (2010) , the result consistently showed that oil price have positive relationship with stock return. However, Chen

et.al (1986), Gay (2006) and Aktham (2004) concluded that oil price cannot predict the stock return.

Table 2.1 Summary Literature Review Author, date Data Gan et al. (2006) Times Series data The sample consist from January 1990 to January 2003 on monthly base Variables Share price Index Inflation rate Exchange rate Gross domestic product Long term interest rate Short term interest Rate Domestic retail oil price

Methodology Unit Root Test Johansen Cointegration Test Granger causality Test Innovation Acconuting Analysis

Finding The Johansen cointegration result shows that there are long run relationship between New Zealand Stock Index and the macroeconomic variables tested. The Granger causality test stated that the New Zealand Stock Index is not a leading indicator in New Zealand. Innovation Accounting Analysis result show that the money supply, inflation rate, exchange rate and long term interest rate have negative impact on the index. However, the result indicates that the gross domestic product and domestic retail oil price have positive impact on index.

Table 2.1 Summary Literature Review (continue) Author, date Data Chen et al. (2005) Hotel Stock are traded in Taiwan Stock Exchange Monthly hotel stock prices from January 1989 to August 2003 Variables included Industrial production Inflation rate Unemployme nt rate Money growth Yield spread Frimpong (2009) Used monthly times series data from December 1990 to December 2006 for Ghana stock exchange Variable Inflation rate Exchange rate Interest rate Money supply

Methodology Regression model

Finding Money supply and unemployment rate have significant to the hotel stock return in Taiwan. The result also show that the nonmacroeconomic have stronger significant to the stock market compare to macroeconomics variables.

unit root test Johansen cointegration test Vector error-correction model

The study showed that the exchange rate has positive effect on Ghana stock market. However, inflation, interest rate, and money supply have negatively significance on Ghana stock market.

Table 2.1 Summary Literature Review (continue) Author, date Data Abugri (2008) Stock Index for four Latin American countries for the period January 1986 to August 2001 on monthly data Variables Included U.S 3 monthly Treasury bill yield MSCI world index Exchange rate Money supply Industrial production Interest rate Eryigit (2009) Identify stock market by using 16 selected sectors in Istanbul stock exchange from 2000 to 2008. Variable Oil Price

Methodology Finding Vector autoregressive (VAR) The MSCI world index and US.3month Treasury-bill yield are model consistently significant for stock market in Argentina, Brazil, Chile and Mexico. The interest rate exchange rates are significance for stock market in Brazil, Mexico, and Argentina. The stock return responds significance to exchange rate in the stock market in Brazil, Mexico. The impact of money supply and industrial production toward stock market in Argentina, Brazil, Chile and Mexico are weak.

OLS regression

The oil prices have positive effect in positive effect on wood, paper & printing, insurance and electricity sub-sector indices. But there are no significance on transportation, banks, leasing, factoring, and real estate investment trust sector.

Table 2.1 Summary Literature Review (continue) Author, date Data Chen et al. ( 1986) The variables computed by monthly from January 1953 through November 1983 Variables Industrial production Unexpected inflation Expected inflation Term structure Oil price Consumption Risk premium Market index Gay (2008) Examine the effect of macroeconomics on stock market return in Brazil, Russia, India and China from 1990 to 2006 Variables Oil price Exchange rate

Methodology Multivariate approach Arbitrage pricing model

Finding Industrial production, risk premium, term structure, unanticipated inflation, expected inflation is significant to stock return. However, the study found that consumption, oil price and market index are not significance to stock return.

Box- Jenkins ARIMA model Dickey-Fuller test

There are not significant effect of exchange rate and oil price on stock return for Brazil, Russia, India, and China.

Table 2.1 Summary Literature Review (continue) Author, date Data Wongbangpo & Monthly data from Sharma(2002) 1985 to 1996 Jakarta composite stock price indek for Indonesia , kuala Lumpur stock exchange composite index for Malaysia ,Philippine stock exchange composite index for Philippines, the stock exchange of Singapore index for Singapore and stock exchange of Thailand index for Thailand. Variables Gross national product Exchange rate Money supply Consumer price index Interest rate

Methodology Cointegration Granger causality Likelihood ration test

Finding There are negative effects of consumer price index on stock price in all countries. Money supply has positive respond on stock price in Malaysia, Singapore and Thailand. The negative relation between money supply and stock price can be observed in Indonesia and Philippine The interest rates have negatively influence stock market in Philippines, Singapore, and Thailand.

Table 2.1 Summary Literature Review (continue)

Author, date Clare & Priestley (1998)

Data Examine 70 stock from Malaysia stock exchange from January 1986 to August 1994 International factor The excess return on the MSWCM index Domestic variables Unexpected change in the risk free rate of interest Unexpected change in the term structure of interest rate Unexpected change in industrial production Unexpected inflation Change in expected inflation

Methodology Non linear seemingly unrelated regression methodology (ITNLSUR)

Finding The unexpected changes in risk free rate, unexpected changes in term structure of interest rates, unexpected inflation, and changes in expected inflation are significant on stock return. However, the industrial production is not effect on the stock return.

Table 2.1 Summary Literature Review (continue)

Author, date Maysami et al. (2004)

Data Investigate the macroeconomics factor of stock return by used selected Singapore exchange sector indicates which including finance index, property index, hotel index from February 1995 to December 2001. Variables Short term interest rate Long term interest rate Inflation rate industrial production Exchange rate Money supply

Methodology Augmented Dickey Fuller Unit root test

Finding The study found that the inflation rate have positive effect on Singapore stock return. The short term interest rate are positively significance to stock return but long term interest rate have negative relation on stock return. Money supply has positive relationship with stock return.

Table 2.1 Summary Literature Review (continue)

Author, date Arouri, et al. (2010)

Data Identify stock return in GCC countries which included Qatar, Oman, Saudi, Arabia, UAE, Bahrain, and Kuwait Variable Oil Price Identify Istanbul Securities Exchange (ISE) from Jun 1991 to March 2000 Variables Foreign exchange rate Growth rate of money stock (currency in circulation) Industrial Production Interest rate

Methodology Jarque-Bera statistic (JB) OLS test Multifactor model

Finding The result shows that oil price are significance positive in Oman, Qatar, Saudi Arabia and UEA stock market. Oil price are not effect in Bahrain and Kuwait stock market.

Erdogan & Ozlale (2003)

Augmented Dickey Fuller test Generalized Autoregression conditional Heteroscedasticity (GARCH)

The study found that before 1994 (financial crisis) the foreign exchange rate have positive effect on stock return. However, the relationship of foreign exchange rate and exchange rate become negative after crisis. Secondly, the industrial production are positive significant to stock return. The relationship have excluded the period from 1994 financial crisis and end with the beginning of 1997 Asian crisis. Thirdly, the interest rates influence the stock return in positive way until 1994 financial crisis.

Table 2.1 Summary Literature Review (continue)

Author, date Rahman et al. (2009)

Data Examine effect of macroeconomics factor in Malaysian stock market. Selected variables Interest rate reserve industrial production Exchange rate Money supply Investigate in AsianPasific which included Malaysia, Korea, Thailand, Hong Kong, Japan and Australia from January 1993 to December 2002. Variables Foreign exchange rate Consumer price index industrial production

Methodology Augmented Dickey Fuller Philip Perron test Kwiatkowski- PhilipSchmidt-Shin Johansen-Jusellus test Likehood ratio Granger causality test

Finding The result shows that industrial production and reserve have positive relationship with stock return. However, the money supply and exchange rate, interest rate has negative effect on stock return.

Mahmood & Dinniah (2009)

maximum , minimum, mean, standard deviation, skewness and kurtosis Augmented Dickey-Fuller (ADF) Engle- Granger cointegration Multivariate Johansen Cointegration test Error correction model

The Engle-Granger test show that Thailand stock return correlate with foreign exchange and inflation while Japanese stock return cointegrated with inflation rate and industrial output. The korea stock return cointegrated with foreign exchange and Hong Kong stock return cointegrated with inflation. The result of ECM model show that only Thailand and Hong Kong have relationship between stock market and macroeconomics variables in the short run.

Table 2.1 Summary Literature Review (continue)

Author, date Gunsel & Cukur(2007)

Data 87 stock which are classified in 10 industry sector and listed in London Stock Exchange for period 1980 to 1993 on monthly base Variables Term structure of interest rate Unanticipated inflation Risk premium Exchange rate Money supply Sector dividend yield Unexpected production

Methodology Durbin-Waltson

Finding The dividend yield has negative effect on the return for all industries. The unexpected inflation only effect on the return of the food, beverage and tobacco industry. The risk premium has closely positive effect on the return for all industries. The exchange rate do not effect on the return for all industries. The money supply may affect one industry positively or negatively. The interest rate have positive relationship with construction, food, beverage & tobacco, oil exploration & production and electronic & electrical equipment industry. The unexpected production affects industry positively or negatively.

Table 2.1 Summary Literature Review (continue)

Author, date Aktham (2004)

Data Selected 22 emerging countries from 1 January 1998 to 31 April 2004 to examine the relationship of oil price and stock return Variables Oil Price Tested 28 industry in Tokyo stock Exchange from January 1973 to December 1998 Variables Money supply Unantipated inflation Industrial production Term structure Exchange rate Growth rate in commercial land price

Azeez & Yonezawa (2006)

Methodology Augmented Dickey-Fuller (ADF) Phillip-Perron (PP) Kwiatowki-Pillips-SchmidtShin (KSPP) VAR analysis Generalized variance decomposition Generalized impulse responses function Likelihood ratio test

Finding The result of generalized decomposition show that crude oil price do not effect on stock return of emerging countries.

The money supply, inflation rate, exchange rate and industrial output are significance to the expected stock return.

Table 2.1 Summary Literature Review (continue)

Author, date Butt et al. (2010)

Data Selected 32 firms from banking and textile sector of Karachi stock exchange. The sample period from Jul 1998 to Jun 2008 Variable Market return Consumer price index Risk free rate of return Growth in industrial production Change in exchange rate Growth in money supply Individual industrial production

Methodology Mean, standard deviation, skewness and kurtosis Augmented Dickey Fuller statistics GARCH techniques

Finding The banking sector are more volatility than textile sector. Money supply almost insignificant on stock return. The money supply has positive relationship with stock return of textile sector. The sector banking show a negative respond on stock return. For the textile industry, consumer price index, risk free rate, and industrial production are negatively but insignificant to stock return. Exchange rate are positively but insignificant to stock return. For banking industry, the market return is positively effect on stock return. Exchange rate, industrial production index, money supply are negatively effect on stock return through the impact is insignificance. In contrast, risk free rate are positively insignificance to stock return.

Table 2.1 Summary Literature Review (continue)

Author, date Bilson et al. (2001)

Data Examined 20 emerging market Sample period from January 1985 to December 1997 Microeconomics variables Political risk measures Trade sector Interest rate Regional market Price to earning ratio Dividend yield Real activity Exchange rate Goods price Money supply

Methodology Multifactor model Dickey Fuller Phillip-Perron Test Correlation test

Finding The exchange rate is significance with the return for twelve markets. However, the money supply, real activity and good prices are not strongly effect on the return on emerging market.

Table 2.1 Summary Literature Review Author, date Data Tursoy et al. Examined 175 stock (2008) which are classified in 11 industry sector and listed in Istanbul Stock Exchange from January 2001 to September 2005 on monthly base Variable Money supply Industrial production Crude oil price Import Export Gold price Exchange rate Interest rate Gross domestic product

Methodology Ordinary least square (OLS) technique

Finding The paper indicates that macroeconomic factor is not effect on stock return in Istanbul Stock Exchange. The paper states that the macroeconomic factor may affect one industry positively or negatively

CHAPTER THREE METHODOLOGY

3.1

Introduction

The study examine the interaction between inflation rate, money supply, industrial production, oil price and stock return in finance sector and trading & service sector. The process of identifies variables, collect data, select timeframe, identify method and test will be conducted to achieve the objective of study. Firstly, the empirical model will be discussed in Section 3.2. Second, the study implement the unit root test to investigate the stationary for all series and Johansen Cointegration test, VECM Granger Causality test, Granger Causality test are conducted to examine the relationship between the series in term of long run and short run. The methodology of study will be further discussed in Section 3.3. The data description for the definition and transformation of variables and time period will be discussion in Section 3.4.

3.2

Empirical Model

The study employs Arbitrage Pricing Theory as an empirical model. The equation can be express as following:

(3.1) (3.2)

58

where

is the natural logarithm stock return on the finance sector and is the constant term and is the reaction coefficient

trading & service sector and

measure the change in return for a change of macroeconomic factors . The set of macroeconomic factors include =Natural logarithm of Inflation rate for period of t = Natural logarithm of money supply for period of t = Natural logarithm of Industrial production for period of t =Natural logarithm of Oil price for period of t

The model (3.1) and (3.2) implies the relationship between the logarithm of inflation rate, logarithm of money supply, logarithm of industrial production , logarithm of oil price and logarithm of stock return in finance sector and trading & service sector. 3.3 3.3.1 Methodology Augmented Dickey-Fuller unit-root Test

The Augmented Dickey- Fuller Unit-root Test is used for examine the stationary of times series data. In ADF test, there are three difference form or difference null hypotheses. For example, the null hypothesis for LGFNSR can be express as following (3.3) (3.4) (3.5)

59

Where

is differencing operator for the series of natural logarithm of stock return is the intercept, and denoted as coefficients, is the times

for finance sector series trend,

is the number of lagged term and

is the error term. Model (3.3) is

pure random, model (3.4) have intercept and lastly model (3.5) have trend and intercept. The hypotheses for the ADF unit roots test are:

The null hypothesis implies that the existence of unit root and the variables is non-stationary. The alternatives hypothesis implies that the variables does not contain unit root which consider that the series is stationary.

The ADF test use t-statistic to identify the stationary of variables. The optimum lag length is chosen on Akaike Information Criterion (AIC). The decision of significance variable is based on critical value in 1%, 5% and lo% significance level. If the t-statistic is greater than critical value, therefore it is significance and reject null hypothesis. In contrast, if the t-statistic is less than critical value, therefore the series is insignificance and failed to reject null hypothesis.

60

3.3.2

Phillip-Perron Test

The PP test is written as (3.6) Where is the times series represented as defined as the innovation term, is

represent the number of observation Similarity with ADF test, the hypothesis for PP test is

. The null hypothesis implies that existence of a unit root. Therefore, the series are considered non-stationary. The alternative hypothesis implies that the series do not contain unit root and the variables is stationary.

PP test implement t-statistic to identify the stationary of series. The bandwidth of PP test is selected using the Newey-West criterion based on Bartlett kernel method. The selection of significance variables is based on the critical value in 1%, 5% and 10% significance level. Similarity with ADF test, the null hypothesis for PP test implies the series contain unit root and the variables is non-stationary. The alternatives hypothesis implies that the variables do not contain unit root and the variables is stationary.

If the t-statistic is greater than critical value, therefore it is significance and reject null hypothesis. The series can be concluded as stationary. However, if the t-

61

statistic is less than critical value, therefore the series is insignificance and failed to reject null hypothesis. It means that the series is non-stationary.

Table 3.1: ADF and PP Critical Value for Monthly Sample Period from January 1998 to December 2010 ADF Critical Value 1% significance 5% significance 10% significance level level level At level With intercept -3.473 -2.880 -2.576 With intercept and -4.019 -3.440 -3.144 trend At first Difference None -2.580 -1.943 -1.615 With intercept -3.473 -2.880 -2.577 With intercept and -4.020 -3.440 -3.144 trend Notes: The critical values are found from E-view 5.

3.3.3

Kwiatkowski-Phillip-Schmidt-Shin Test

The KPSS test is adapted in the following regression (3.7) Where defined as partial sum of residual, defines as number of observation. represent error variance, , is

the lag parameter and

The hypotheses for KPSS test are =0

62

Distinguish with ADF test and PP test, the null hypothesis for KPSS is the series do not contain unit root and the variables is considered as stationary. In contrast, the alternatives hypothesis is the series contain unit root and variables is non-stationary . The KPSS test use t-statistic to examine the stationary of series. The bandwidth is chosen using Newey-west criterion based on Bartlett Kernal method. If the t-statistic is greater than critical value, the series is significance and reject null hypothesis. Therefore, the variables can be concluded as stationary. In contrast, if the t-statistic is less than critical value, the series is insignificance and failed to reject null hypothesis. Thus, the variable is non-stationary.

Table 3.2: KPSS Critical Value for Monthly Sample Period from January 1998 to December 2010 KPSS Critical Value 1% significance 5% significance 10% significance level level level At level With intercept 0.277 0.463 0.347 With intercept and 0.216 0.146 0.119 trend At first Difference With intercept 0.739 0.463 0.347 With trend and 0.216 0.146 0.119 intercept Notes : The critical value are found from E-view 5.

63

3.3.4

Johansen Cointegration Test

After identify the stationary of the variables, the Johansen Cointegration Test is conducted to examine the long run relationship between the stock return and macro variables. The study is allowed to conduct the test if all the variables is stationary at level I(0) or I(1). The Johansen Cointegration test can be referred as following regression

(3.8) where for i = 1,2,K-1; is an identity matrix The denoted as short term adjustment parameters and comprise long-term

equilibrium relationship X variables. matrix and . Therefore .

decomposed into the product of two n by r is a matrix contain r cointegration vectors and

denoted as the speed of adjustment parameter.

Two likelihood test use for examine the number of cointegration vector. The likelihood test consists of trace test and maximum Eigenvalue test. The trace test is described in the following regression of; (3.9) Where T is the number of observation, N is number of variables, and estimated eigenvalue. is the largest

64

R=0 R 0

The null hypothesis implies that there is no cointegration and alternative hypothesis implies that there is one cointegration vector. If the computed value of the trace test is larger than the critical value. It indicates that reject null hypothesis and this conclude that the model has cointegration model. In contrast, if the computed value of the trace test is less than the critical value therefore it is fail to reject null hypothesis. Thus, the model has no cointegration model.

The Eigenvalue Test is expressed as following;

(3.10) where T is the number of observation and hypothesis under Eigenvalue Test is cointegrating vector cointegrating vector The null hypothesis indicates that there is no cointegration factor in the long run and the alternative hypothesis defines as there is cointegration factor in the long run. When the computed value of Eigenvalue test is greater than critical value, it means that reject null hypothesis. Therefore, the model has cointegration factor in the long run. However, if is the largest estimated eigenvalue . The

65

the computed value of Eigenvalue test is less than the critical value, it mean that fail to reject null hypothesis. Therefore, the model has no cointegration factor.

3.3.5 Vector Error Correction (VECM) Model Granger Causality

Once the cointegration is identify in the Johansen Cointegration test, the Vector Error Correction (VECM) model Granger Causality test is used for examine the short run relationship between the variables. In case of trading & service sector, the regression is express as below

(3.11)

(3.12)

(3.13)

66

(3.14)

(3.15) where is the error correction term and are parameters , K is

the lag length, variance.

is the stationary random process with mean zero and constant

In the model (3.11), (3.12), (3.13), (3.14) and (3.15), the null hypothesis is . It indicate that the stock return do not Granger-cause selected variables. In contract, alternatives hypothesis conclude that stock return for trading & service sector have granger cause those selected variables.

If the variables are statistically significance at 1% , 5% or 10% level where probability is less than the significance level, it mean that reject null hypothesis. Thus, it can conclude that there are granger causes between the variables. However, if the variables are insignificance at 1%, 5% or 10% level where the probability is greater

67

than the significance level, it is fail to reject null hypothesis. Therefore, it conclude that those variables do not granger cause each other.

3.3.6

Granger Causality Test

Once the result of Johansen Cointegration Test show that there is no cointegration in the long run, the Granger Causality test is conducted for identify the short run relationship between the variables.

The null hypothesis and alternative hypothesis for the finance sector are represented in Tables 3.3. If the variables are significant at 1%, 5%, 10% level where pvalue is smaller than the significance level, the null hypothesis is rejected. Therefore, it can conclude the independent variables cause the dependent variables. In contrast, if the variables is insignificance at 1%, 5% or 10% where p-value is greater than the significance level , the null hypothesis is failed to reject. It concludes that there are not causality linkage between the independent variable and dependent variables.

68

Table 3.3: The Null and Alternatives Hypotheses of Granger Causality Alternative Hypothesis, Null Hypothesis, Dependent Variables =LGFNSR LGINF does not Granger Causes LGFNSR LGINF does Granger Cause LGFNSR LGIP does not Granger Cause LGFNSR LGIP does Granger Cause LGFNSR LGMP does not Granger Cause LGFNSR LGMP does Granger Cause LGFNSR LGOP does not Granger Cause LGFNSR LGOP does Granger Cause LGFNSR Dependent Variables =LGINF LGFNSR does not Granger Causes LGINF LGFNSR does Granger Cause LGINF LGIP does not Granger Causes LGINF LGIP does Granger Causes LGINF LGMP does not Granger Causes LGINF LGMP does Granger Causes LGINF LGOP does not Granger Causes LGINF LGOP does not Granger Causes LGINF Dependent Variables =LGIP LGFNSR does not Granger Cause LGIP LGFNSR does Granger Cause LGIP LGINF does not Granger Cause LGIP LGINF does Granger Cause LGIP LGMP does not Granger Cause LGIP LGMP does Granger Cause LGIP LGOP does not Granger Cause LGIP LGOP does Granger Cause LGIP Dependent Variables =LGMP LGFNSR does not Granger Cause LGMP LGFNSR does Granger Cause LGMP LGINF does not Granger Cause LGMP LGINF does Granger Cause LGMP LGIP does not Granger Cause LGMP LGIP does Granger Cause LGMP LGOP does not Granger Cause LGMP LGOP does not Granger Cause LGMP Dependent Variables = LGOP LGFNSR does not Granger Cause LGOP LGFNSR does Granger Cause LGOP LGINF does not Granger Cause LGOP LGINF does Granger Cause LGOP Null Hypothesis, LGIP does not Granger Cause LGOP LGMP does not Granger Cause LGOP Alternative Hypothesis, LGIP does Granger Cause LGOP LGMP does Granger Cause LGOP

3.4

Data Description

The data for the variables obtain for the period of 13 year which starting from January 1998 to December 2010. The reason for selecting the time period is because the data of stock price for mostly companies are not available in 1997. Mostly, the

69

companies have not listed during the period. The monthly data is chosen because the variables are available monthly in Malaysia.

In the study, Malaysia stock market will be chosen to investigate the relationship between macroeconomics factors and stock return. The reason for selecting Malaysia as proxies for underlying risk is because Malaysia consider as emerging country. Malaysia has a constant growth in Gross Domestic Product. Therefore, many investors are interested to invest capital in Malaysia.

Since the objective of study is to determine the relationship of macro variables and stock return by sector approach, therefore there are two main sectors will be chosen in the study. The two main sectors are finance sector and trading & service sector. The reason for finance sector and trading & service sector are selected in this study is because the both sector have contributed high GDP growth in Malaysia. According to Ministry of finance (2010) in economic report 2010/2011, the finance sector contributes RM25, 645 million in year 2010 in the fourth quarterly which is higher than other sector such as construction sector, storage & communication sector. Besides that, the developed Islamic financing in finance sector have increased interest among the public. For the trading & service sector, the share of GDP in Malaysia was 57.6% in 2010 (Ministry of finance, 2010). The trading & service sector is the main sector for contribution GDP in Malaysia. Therefore, it is important to understanding and examine the trading & Service sector and finance sector in stock market.

70

The study consists of 107 companies in finance sectors and trading & service sector. 32 Out of 107 companies are listed in finance sector and 75 companies listed in trading & service sector. The inflation rate, industrial production, money supply, oil price and stock return are employed to investigate the effect of stock return. Those selected variables have been by previous studies. The Tables 3.4 show the macroeconomic that are employed in literature studies.

Table 3.4: Macroeconomics Variables That Are Implemented in Previous Studies Macroeconomics Previous studies that employ indicated Variables variables Inflation Rate Gan et.al (2006), Chen et.al (2005), Frimpong (2009), Wongbangpo & Sharma (2002),Clare & Priestley (1998), Maysami et.al (2004), Mahmood & Dinniah (2009), Industrial Chen et.al (2005), Aburgi (2008), Clare & Priestley (1998), Production Maysami et.al (2004) , Erdogan & Ozlale, Rahman et.al (2009), Mahmood & Dinniah (2009), Azeez & Yonezawa (2006), Tursoy et.al (2008) Money Supply Chen et.al (2005), Frimpong (2009), Aburgi (2008), Wongbangpo & Sharma (2002), Maysami et.al (2004), rahman et.al ( 2009), Gunsel & Cukur (2007), Azeez & Yonezawa (2006), Tursoy et.al (2008) Oil Price Eryigit (2009), Gay (2008),Aroubi et.al (2010), Aktham (2004), Tursoy et.al (2008)

71

The Definition of the variables is described in Table 3.5 Table 3.5: Definition and Transformation of Variables Variables Transformations Definition of variables Definition of Transformation Natural logarithm of weighted average of monthend closing prices for finance company shares listed on KLCI index. Monthly return on the Finance Index Natural logarithm of weighted average of monthend closing prices for trading & service company shares listed on KLCI index. Monthly return on the trading & service index Natural logarithms of the month-end consumer price index Monthly-realized inflation rate Natural logarithms of the month-end industrial production index Monthly growth rate of industrial production Natural logarithms of the month-end industrial production index Monthly growth rate of money supply Natural logarithms of month-end oil price in term US dollar Monthly growth rate of oil price The money supply is measures by M2. The industrial production is measures by output. Consumer price index is used to represent inflation rate. The domestic retail oil price in term of US dollar is used to measure of oil price. Data for money supply, consumer price index, oil price and industrial production are drawn from Bank Negara Malaysia. The data for stock prices are available at Datastream database

72