Stock Prices Movements Have Always

Published Date: 02 Nov 2017

1.0 Introduction

Stock prices movements have always been an interest of study to researchers as well as investors. Researchers explore the movements of stock prices to determine its behaviour and pattern as well as the underlying forces while investors observe stock price movements to evaluate capital gains and profits.

Generally, stock prices tend to follow random walk movement, that is, the movements of stock prices are random and cannot be predicted, especially in an efficient market, because the prices reflect and respond quickly and accurately to relevant information, as proposed in the Efficient Market Hypothesis (EMH) and this random information is the cause of the future stock price changes (Lee & Lee, 2009). Nonetheless, some of the information can be obtained to forecast and estimate the future price movements. Macroeconomic variables are one of the factors will affect the stock price changes.

In this study, we are going to investigate how macroeconomic variables affect stock prices of financial sector in Malaysia. Chapter 1 includes research background, problem statement, research objectives, research questions, hypothesis of the study, significance of study, chapter layout and concluding section.

Research background

History and Background of Bursa Malaysia

In Malaysia, all the stock prices are displayed and recorded at Malaysian stock exchange market, which is now called Bursa Malaysia. Bursa Malaysia is the exchange holding company which regulates the trading activities in Malaysia. Initially, the Bursa Malaysia is called Singapore Stockbrokers’ Association which established in 1930, but re-registered as Malayan Stockbrokers’ Association in 1937. In year 1960, the Malayan Stock Exchange was established and the purpose is public trading of shares commenced. On April 14, 2004, the name was changed from Kuala Lumpur Stock Exchange Berhad to Bursa Malaysia Berhad, the purpose is to enhance the competitive position and to global trends in exchange sectors by making us have more customer. In the following year, Bursa Malaysia was listed on the Main Board of Bursa Malaysia Securities Berhad.

There are many types of sectors contributing to economy such as manufacturing, investment and finance, industrial, consumer product, agricultural and many more. The stock price of each market is shown in the Main Market in Bursa Malaysia. Now there are altogether 929 companies listed in Bursa Malaysia with 814 in Main Market and 115 in ACE market ("ACE" is short for "Access, Certainty, Efficiency") (Bursa Malaysia-Listing Statistic). These listed companies play a vital role in the economy because stock prices can actually indirectly reflect the economy of a country (Pearce, 1983). An average low-perform stock market indicate that the country is undergoing a recession period.

Importance of Investment and Finance Sector

One of the sectors listed on Bursa Malaysia is the investment and finance sector. This sector is considered one of the most important sectors that make up Malaysian market (Ahmad, 2005). In Malaysia, there are 39 listed companies in finance sector of Malaysia comprising of commercial banks, such as Malayan Bank (or Maybank) and Public Bank, investment banks, such as OSK Investment Bhd., Islamic banks such as Bank Islam as well as insurance companies such as Malaysian Assurance Alliance Berhad and other financial service companies like Aeon Credit Service.

Financial sector plays a very crucial role in facilitating the channel of funds nationwide and worldwide as an intermediation function and an entity of economic growth promotion (Jalloh, 2009; Yu & Gan 2010). Capital markets and financial intermediaries help in mobilizing resources, assisting business and individual entities to manage their risk and return as well as monitoring investment projects (Gimet & Lagoarde-Segot, 2012). The researchers also suggested that financial access could help in mitigating poverty as well as income inequality. Besides, the relationship between the financial sector development and economic development is an interesting subject of study among researchers (Yu & Gan, 2010), such interested researchers include Demetriades and Hussein (1996), Arestis, Demetriades and Luintel (2001) as well as Levine (1997).

Methods of Forecasting Stock Prices Movements

Investors pour their money in to financial investment vehicles to hope that they would get a return as a reward for delaying consumption. As rational investors, they would prefer high returns and therefore will buy at a price as low as possible and then sell it at a perceived peak price. In order to get high return, many would predict the stock price through various methods such as looking for trends and cycles and econometric methods such as autoregressive integrated moving average model (ARIMA model in short). If the publicly available information regarding on the stock prices and relevant information can be predicted for stock return, it is actually violating the principle of semistrong market efficiency (Chen, 2009). Effecient Market Hypothesis (EMH) postulates that all relevant information currently known about changes in macroeconomic variables are already fully reflected in current stock prices and because of the investors competition, abnormal returns are ceased to exist through future stock market movements prediction (Zahid Ahmad, 2012).

Other than those methods, macroeconomic variables also proved to be a very suitable method. There are many economists and finance specialist interested in investigating the macroeconomic variables affecting the stock price. According to Chen (2009), macroeconomic variables play a very important role in determining aggregate stock market behavior. Some of the macroeconomic variables include money supply, inflation rate and unemployment rate. Most of the existing empirical researches on the stock price focus on the relationship between macroeconomics variables and the overall performance of stock price in the economy (Black, Fraser, & Groenewold, 2003; Mohammad, Hussain, Jalil, & Ali, 2009; Yahyazadehfar & Babaie, 2012). And thus, the empirical results only showed a general description across the industries.

1.2 Problem statement

In this paper, we are attempting to estimate the relationship between macroeconomics variables affecting the stock prices of financial sector in Malaysia, which is more specific and interesting in a manner.

Malaysian stock market is an interesting topic to study upon because of its unique features. First, Malaysian stock market is of not a weak-form efficient and it shows significant predictable behaviour in the stock market and thus the stock market does not follow random-walk theory (Hafiz, Jae, Chong, 2007). Second, a lot of studies examining the determinants of stock prices primarily focus on well-developed markets and little attention is paid on the emerging markets (Humpe & Macmillan, 2007; Hardouvelis, 1987). Third, not much studies have been done based on particular economic sectors as most of the studies examined the overall market performance. This paper will, thus, examine the performance of the stock market of Malaysia when there is a variation in the macroeconomic variables of Malaysia. In addition, we will narrow down our investigation scope to only the financial sector in Malaysia, but not the 100 largest companies in Malaysia that make up the Kuala Lumpur Composite Index (KLCI).

Research objectives

General objective

This paper aims to study how macroeconomic variables affect the movement of stock prices of financial sector in Malaysia.

Specific objectives

The specific objectives are as below:

To investigate whether there is a long run relationship between the macroeconomic variables examined and the stock prices of financial sector in Malaysia.

To examine the significance of relationship between exchange rate and stock price of financial sector in Malaysia.

To examine the significance of relationship between inflation rate and stock prices of financial sector in Malaysia.

To examine the significance of relationship between money supply and stock price of financial sector in Malaysia.

To investigate whether causal unidirectional or bidirectional effects exist between the macroeconomic variables and stock prices of financial sector in Malaysia.

Research questions

In this research, we will attempt to find out:

How macroeconomic variables affect stock price of financial sector in Malaysia in a long run?

Are the macroeconomic variables significant in explaining the stock price movements of financial sector in Malaysia?

Will there be an existing of causal unidirectional or bidirectional effect of macroeconomics variables on stock price of financial sector in Malaysia?

Hypothesis the study

Based on the reviews stated above, it is confirmed that macroeconomic variables are important in explaining the stock price movements, but yet to know what relationships may exist between the macroeconomic variables and stock prices.

Therefore, our preliminary hypotheses of our study are:

Ho: None of the macroeconomic variables is significant to the stock price of financial sector in Malaysia in a long run.

H1: At least one of the macroeconomic variables is significant to the stock prices of financial sector in Malaysia in a long run.

Significance of the study

We believe that after conducting this study, it can enrich our understanding of how financial sector responds to macroeconomic changes in Malaysia, and therefore affecting the entire financial system of Malaysia as a whole. For policymakers, the study provides reliable information and some in-depth insights on the sensitivity of financial sector towards the changes of macroeconomic variables and we hope that the policymakers would take extra precautions in regulating Malaysian macroeconomic factors. Besides, it also can be a reference for investors to monitor their invested stock price movements whenever there is a significant change in macroeconomic environment, for example: high inflation rate, depreciated home currency (Malaysian Ringgit) or low gross domestic product growth.

Chapter layout

The chapter layout in this paper is as follow: Chapter 1 provides the objectives of study and introduces the research purpose and the reason we choose financial sector as our target data for research. Chapter 2 contains a comprehensive literature review; Chapter 3 explains the research methodology used for the analysis of this paper; Chapter 4 consists of the analysis of data and empirical results while Chapter 5 includes our discussions, conclusion and suggestions of policy implication.

1.8 Conclusion

As the existence of relationship between macroeconomic variables and stock prices has been revealed, a more focused research is conducted to examine its hidden mechanics in a long run, especially in Malaysia.

Throughout this research, interesting findings and implications have been obtained and we will discover the elements from this research that differs from other research done by previous researchers from the literature review and provide comprehensive results from our investigation on macroeconomic variables effects on the stock prices of the financial sectors in Malaysia.

CHAPTER 2: LITERATURE REVIEW

Introduction

Many studies have been conducted to establish the relationship between stock price and the macroeconomic variables. In this literature review, we had summarized some important studies in regard. The dependent variable of the research is stock price of financial sector listed on Bursa Malaysia. The independent variables are money supply, real exchange rate denominated in MYR (Malaysia Ringgit) against USD (United States Dollar) and inflation rate. The objectives of this chapter are to provide supporting details for the selected macroeconomic variables and to understand how each independent variable affect dependent variable as well as critical review on past researchers’ work on similar topics. The organization of literature review is as followed: review of literature, review of relevant theoretical models, proposed theoretical/conceptual framework, hypotheses development and conclusion.

Review of the Literature

Review on Methodology

Stock prices play an important role in every country in term of economy. Thus, there are some literatures in investigating the relationship between stock market and macroeconomic variables. The different of macroeconomic variables in different countries and different econometric methods as well as study period may affect the result stock market price. For example, Maysami and Koh (2000) applied Johansen’s VECM to analyze the relationship between the Singapore Market and inflation, money supply growth, changes in short- and long-term interest rate and variations in exchange rate. They concluded the dependent variable and independent variables formed a cointegrating relationship with changes in stock market levels.

Furthermore, Hassan (2003) employed Johansen and Juselius’ multivariate cointegration techniques to test for the existence of long-term relationships between share prices in the Persian Gulf region. Using a vector-error-correction model, he also investigated the short-term dynamics of prices by testing for the existence and direction of inter-temporal Granger-causality.

Besides, Oyama (1997) investigated the phenomenon of the determination of US stock price by means of macroeconomic variables (money growth rates (M1, M2), inflation, interest rate, exchange rate, commodity price index). He used error correction model, multi-factor return generating model and revised dividend discount model to draw conclusion empirically. Husain & Mahmood (2001) examined the general relationship between macroeconomic variables (economic activity (GDP), investment spending and consumption expenditure) and stock price in developing economy of Pakistan from 1960 to 1999 as study period. Cointegration analysis and error correction estimation techniques were applied to investigate the long run relationship.

Overall, the techniques used to investigate the relationships are cointegration, error correction model and Granger causality. Thus it can be said that these are the fundamental tests required for such research, though there are still no research done in explaining the effect on stock price of any sector in Malaysia by macroeconomic variables. Hence, this paper attempts to select several macroeconomic variables such as exchange rate, inflation (consumer price index) and money supply (M2) to examine the hypothesized relations with stock price index of financial sector in Malaysia.

Review on Money supply

According to the IMF’s manual, money supply is measured as the combined deposit liabilities of the banking system and the currency liabilities of the central bank, both held by households, firms, nonprofit institutions and all public sector entities outside of the central government.

Homa and Jaffee (1971) analyzed the relationship between money supply and stock prices by using quarterly data of the growth rate of money supply and Standard & Poor’s 500 index as proxy of stock prices. The result showed that the two variables are significantly correlated.

Husain and Mahmood (1999) tested the causal relationship between stock returns and money supply in Pakistan. The authors used data of money supply (M2) and six stock price indices (one general and five of different sectors) with time period of June 1991 to June 1999. To investigate the relationship between variables, cointegration analysis and error correction model were implied. A long run relationship was found between stock prices and money supply by means of cointegration analysis. For causality test, a unidirectional relationship was found running from M2 towards stock price. Chandran & Norazman (2004) investigated the causality between money supply and stock prices by using simple bivariate Granger causality test for Malaysia stock market. The result indicates that there is a strong existence of bi-directional relationship between money supply and stock prices.

The competing theories to be examined the relationship between money supply and stock price are the ones developed by the Keynesian economists and the real activity theorists (Sellin, 2001). Keynesian economists argue that there is a negative relationship between stock prices and money supply. According to them, an increase of money supply will lead people to predict tightening monetary policy in the future. Hence, they bid for funds for their prediction, which will drive up the current rate of interest. The interest rates as well as the discount rates go up and the present value of future earnings falls. Stock prices consequently decrease. Besides, other arguments include Ahmad, Ahmad, Khan, & Javaid and Gan, Lee, Yong, & Zhang (2006), who said an increase in money supply caused an increase in inflation, thus people may want to keep real cash and increase their spending expenses hence people sells their shares and other security rather than involving in risky investment. This causes a decrease in prices of securities. The result is further supported by researchers such as Mohammad et al. (2009) and Humpe & Macmillan (2007).

Whereas real activity theorists argued that there is positive relationship for the two variables. The economists believed that increase in money supply means that money demand is increasing in anticipation of increase in economic activity. Higher economic activity implies higher expected profitability, which causes stock prices to rise (Sellin, 2001). Sellin (2001) argued as money supply increases, it actually reflects money demand is increasing, which gives signals of an increase in economic activity which in turn increases cash flows and causing stock prices to rise. This is consistent with Sohail & Hussain (2009) and Maysami & Koh (2000) found a positive relationship between the money supply and stock prices further support this hypothesis.

Thus, the direction of the impact of money supply on the stock prices is not clear although many studies provided evidence of a strong relationship between these two variables (Nishat & Shaheen, 2004)

Some other reasons why there are two types of relationship exist between money supply and stock price may because there is a possibility of different method, country and time period used for the research. For example, Sohail & Hussain (2009) and Mohammad et al. (2009) studies the relationship between money supply and stock price in Pakistan but they obtained different results as they used different test to study the models. The former used vector error correction model (VECM) while latter applied Auto Regressive Integrated Moving Average (ARIMA) to detect the correlation between money supply and stock price in Pakistan.

Exchange rate

Exchange rate can be defined as the nominal exchange rate that takes into account with the inflation differential among the countries (K.A.N and K.M, 1997). Based on traditional approach, changes of exchange rate may lead to some impact in Malaysia stock prices due to the demand and supply effects by investors (Aydemir & Demirhan, 2009). Besides that, economic conditions such as crisis may cause a change in country exchange rates and lead to changes in stock prices.

Economic condition with crisis and without crisis will cause the investors’ purchasing power to increase or decrease and hence influencing the demand and supply of stock market as well as stock prices. According by Pan, Fok, & Liu(2007), and Mohamed, Wisam, Aris, & Fouad (2010), the empirical results showed the exchange rate and stock prices for Hong Kong, Japan, Malaysia, and Thailand before the 1997 Asian financial crisis have a significant causal relationships but it shows no significant causality from exchange rate to stock price during crisis except Malaysia. From the research conducted by Mohamed, Wisma, Aris, & Fouad (2010), it is found that stock prices and exchange rate have a negative casual relationship after crisis, while movement of stock price and exchange rate does not have long run relationship for pre and post crisis (Baharom, Royfaizal, & Habibullah, 2008).

The research by Yau & Nieh (2009) about the stock prices of the exchange rate effects of New Taiwan dollar against Japanese Yen (NTD/ JPY) from January 1991 to March 2008 found a relationship existed between stock prices and exchange rate by having a strong cointegration between NTD/JPY and NTD/USD in a long run instead of short run. Yau & Nieh (2009) found the exchange rate of Taiwan Dollar (NTD) and Japanese Yen (JPN) have highly cointegrated between each other is because the volatility of NTD/JPN will affect the export and import of the Taiwan. As an export-oriented country of Taiwan, the appreciated of NTD against JPY will cause the stock prices to decline. This is due to Taiwan exported firms may lose competitive advantages in world market and causes their stock prices fall.

Lean, Narayan, & Smyth (2011) has their empirical result showing there is little evidence in among Asia markets between the exchange rate and stock prices while Korea is only one of the countries with a weak long-run unidirectional Granger causality which will be affected by the fluctuation of exchange rate toward stock price.

Aydemir & Demirhan (2009) investigated the exchange rate and all stock market indices. They concluded that the results tend to have bi-directional causal relationship, which means the exchange rate can affect the stock market indices and the stock market indices can also affect the exchange rate. Rahman & Uddin (2009), however, found out that there is no way causal relationship can occur between stock prices and exchange rates, which means stock prices do not cause exchange rate and exchange rate does not have an effect on stock prices.

During 1997-8 financial crisis, Malaysia ringgit was pegged at MYR3.80 against US dollar. This represented the depreciation from pre-crisis peak. In order to reduce loss of profit; shareholders will liquidate the shares holding as currency to gain back their cash. Therefore, overall, Malaysia stock market declines because there is a decrease of capital flow from foreign and domestic investors (Azman-Saini, Habibullah, Law, & Dayang-Afizzah, 2006). Changes of the exchange rate after crisis would also affect the firms competitive advantages such that the devaluation of the local currency will lead to foreigner’s demand of local goods and services increases and thus increases of local firm profits and subsequently increases of firm stock price (Mohamed, Wisam, Aris, & Md, F., 2009). However, from the research conducted by Baharom, Royfaizal, & Habibullah (2008), movement of exchange rate and stock prices does not have relationship before or after crisis is because the stock market is assumed to be an efficient market which means the stock prices reflect all available information and will adjust quickly to new information. Therefore, investors are unable to predict the movement of the growth of exchange rate toward stock prices in order to earn abnormal return because of symmetric information in public.

Little evidence among Asia countries found by the researches is because the market transmission information is efficient (reference needed). Weak long-run unidirectional Granger causality relationship between exchange rate and stock market in Korea shows a consistent view with the traditional view. This reveals that intervention of Korea government toward exchange rate could be used to stabilize stock prices (Lean, Narayan, & Smyth, 2011)

Inflation rate

Rise in price level of goods and services will cause the purchasing power and value of money to decline. This is called inflation. Normally it arises from demand pull inflation and cost push inflation that can be controlled using fiscal policy, monetary policy, and supply side economies policy (Riley and College, 2006). Change in stock prices is derived from impact of inflation, such as:

1) During low inflation, cost incurred within a company will drop, thus profit earned will rise;

2) During high inflation, cost incurred within a company will rise, thus profit earned will drop (Graham, n.d.).

The stock prices can be sensitive to inflation in the short term and medium term due to inflationary shocks may have light long term effect on real stock returns (Valcarcel, 2012). The movement of inflation is significant in determining the level of stock price, in which we are conducting this research to prove the relationship between inflation and stock price in the long run.

Many researchers found out there is a negative relationship (Cohn & Lessard, 1981; Ritter & Warr, 2002; Maghayereh, 2003; Boucher, 2006; Sohail & Hussain, 2009) and weakly negative long run relationship (Valcarcel, 2012) between the inflation rate and stock prices. The studies suggests that decrease in after tax profits (Cohn & Lessard, 1981) and unable to adjust firm’s earning based on depreciation of nominal liabilities (Ritter & Warr, 2002) will cause stock price to decrease during high inflation.

Boucher (2006) concentrated on the subjective inflation risk premium explanation. The researcher considered a present value model with a conditional time-varying risk premium to determine long-term trend between both price-earnings ratio and actual inflation. He found out that the transitory deviations are inability to forecast stock returns within short and medium time period, especially in post-war period.

As for Valcarcel (2012), the uncertainty in the correlation between stock prices and inflation arise from two disturbances:

Idiosyncratic stock market shocks

Idiosyncratic stock market increases demand for equities or liquidity that causing the stock prices and US inflation to have negative relationship.

Inflationary or money supply shocks

Inflationary shocks will lead to expansion of money supply that put downward pressure on interest rate, in which directly causes stock prices and US inflation to have positive relationship. After all, the outcomes from both disturbances manage to offset each other to have weak negative relationship.

However, the findings are opposed by Mohamed, Wisam, Aris & Md (2009), Gregoriou & Kontonikas (2010) and Alagidede & Panagiotidis (2010). They discovered that stock price and inflation have positive long run relationship, in which stock market is able to provide hedge against inflation. Mohamed, Wisam, Aris & Md (2009) further investigated the bi-directional relationship between stock price and inflation during pre- and post-financial crisis 1997-8. They found out that stock price and inflation may indirectly have bi-directional relationship because inflation is able to influence exchange rate while exchange rate and stock price has bidirectional causality, on pre-crisis. For post-crisis, stock price and inflation has no causality. Alagidede & Panagiotidis (2010) found out stock price and inflation has insignificant negative relationship in the short run and significant positive relationship in the long run for Egypt and South Africa. Throughout the time period, Nigeria has the significant positive relationship, and Kenya and Tunisia have insignificant positive relationship.

Gregoriou and Kontonikas (2010) used data across different sub-periods, ranging from 1970-1979, 1980-1989, and 1990-2006, in order to test in different inflationary environment. In 1970s, global economy faced two oil shocks that caused massive inflation occurs; in 1980s, the inflation started to move moderately; in 1990s and 2000s, the inflation was under-controlled. The researchers found out that the price elasticity of stocks increased over time when inflation move from a high to a low regime, and low inflation countries and inflation targeting countries are facing greater estimated elasticity.

The difference in results may be rooted from the models the researchers based upon. The research conducted by Cohn and Lessard (1981) and Ritter and Warr (2002) are based on modification of Modigliani-Cohn Hypothesis. They amended the biasness on the hypothesis to come out with the result that stock price and inflation are negatively correlated. As for Alagidede and Panagiotidis (2010), they had achieved positive relationship of stock price and inflation because they used Fisher hypothesis on determining Africa’s stock market.

Increase in inflation will lower down forecasted real earnings growth that lead to decrease in price-earnings ratio and increase in expected return (Boucher, 2006). Campbell (2004) indicates that inflation rate is the main determinant of nominal bond yields which is widely used for years, in which the Fed model also suggest stock yield and inflation are highly correlated. In another word, inflation is also an important factor in explaining our model because it also deals with financial instrument.

Review of Relevant Theoretical Models

In examining stock prices as well as stock return, there are several theoretical backgrounds to support the changes as well as the nature of the stock price. One of such important models frequently used by past researchers is the Capital Asset Pricing Model (CAPM) to determine the required rate of return. This theory relates the expected price of an asset to its riskiness measured by the variance of the asset’s historical rate of return relative to its asset class (Sharpe, 1964). This model assumes that investors choose their portfolio based on Markowitz mean-variance criterion. This model concludes the effect of all factors on stock market, whereas the stock market may be affected by several factors. These can be divided into macroeconomic and microeconomic factors including economic development, gold price, house price, money supply, interest rate, exchange rate and internal actors of companies and economic institutes such as profit division, company agenda and so on (Yahyazadehfar & Babaie, 2012). The equation of the CAPM is therefore:

Ki = Krf + βi (Km – Krf)

Where,

Ki = the required return on security i;

Krf = the risk-free rate of interest;

Km = the return on the market index;

βi = the beta of security I;

(Km – Krf) is known as market risk premium or risk premium.

βi (Km – Krf) is known as asset risk premium

The researchers who employ this pricing model in their research include Kuwornu & Owusu-Nantwi (2011) and Stapleton & Subrahmanyam (1983).

Stock prices and return, no doubt, can be estimated using the said CAPM. But therei s one more issue needed to be address. It is the hypothesis that held researchers for so long in explaining the stock price movements. And that one hypothesis has been postulated to state that stock prices are unpredictable and the market should be efficient as any available information will be incorporated within the stock prices. This hypothesis is the Efficient Market Hypothesis.

The efficient market describes the market price that fully reflects all available information was coined by Fama (1970). Menike (2006) stated the importance of the capital market efficiency is the investors have to base capital market efficiency their investment decisions on information. The investor considers the accurate of the information when evaluating the information. The random-walk hypothesis is based on efficient market assumption that investors react rapidly to new information available. The stock price cannot be predicted because it does not follow a noted pattern. The theory assumes that stock prices are essentially random and therefore, there is no abnormal profit in the stock market. The macroeconomic approach argues that stock prices are sensitive to changes in macroeconomic variables (Abraham, n.d.). Furthermore, Fama (1970) classifies the market efficiency into three levels of the information which is weak, semi-strong and strong forms. Hadi (2006) stated that the weak-form occurs when the stock prices reflect information about the past share price series only. The market is efficient in semi strong form if the security prices reflect not only the information that contains the past time series of stock prices but also all publicly available information. The strong form of market efficiency occurs if the stock price reflects all public and private information. But it is difficult to test strong because private information is difficult to observe. This hypothesis has been the base of several researchers’ research, for example: Hadi (2006), Fama (1991) and LeRoy (1989).

CAPM and EMH seem to be contradicting each other, as one tries to explain the asset prices reasonably and predictably while the other refutes the predictability of asset prices. The key concept in adopting either one of these frameworks is by defining the objectives of study. For our research, we are interested in finding the significance of relationships between the macroeconomic variables (independent variables) and stock prices of financial sector in Malaysia (dependent variable). Thus, we are estimating the model and explaining it in a meaningful sense. Therefore we will not base our research entirely on EMH.

Proposed theoretical / Conceptual framework

An alternative framework to the CAPM is the arbitrage pricing model (APM), in which it is a general theory of asset pricing that has become significant in the pricing of assets. In this research, we will base on this framework to run our tests. This theory was developed primarily by the economist Stephen Ross in 1976 as an alternative to the CAPM. It is a multi-factor model in which every investor believes that the stochastic properties of returns of capital assets are consistent with factors structure (Kuwornu & Owusu-Nantwi, 2011). As argued by Ferson & Harvey (1998) the CAPM and APM have advantages and disadvantages as models of asset prices.

The CAPM is seen as commonly used by equity analysts, but it requires a precise identification of the portfolio against which the asset is compared. While the APM includes multiple sources of risk and alternative investments, suffers from a similar challenge of identification since many factors (international and domestic) that could affect an asset’s performance (Mosley & Singer, 2008). In other words, the independent variable side suffers from the precise number of coefficients needed to explain the model properly.

Nonetheless, this research will employ the asset pricing model as its theoretical framework where the all share index represent the asset price and exchange rates, inflation rates and interest rates will be viewed as their associated risk factors, Abraham (n.d.). Other researchers who adopt the APM are Ahmad et al (2012). Abraham’s (n.d.) research applied APM in which his model is as follow:

Rt = α + β1X1t + β2X2t+… βkXkt + µ

Where,

Xit is risk factor

In our research, we would apply and modify APM since it can include multiple risks (which we would assume that the macroeconomic variables are the risks that would affect the stock prices) and the stock return would be a rough estimate of the stock price, and it would be much easier to use. Our variables are listed as below:

Dependent variable

Independent variables

Stock prices of finance industry

Money supply

Exchange rate

Inflation rate

Table 2.1: Factors affecting stock prices in Malaysia financial industry

The importance of the money supply as a determinant of stock prices may be derived both from the structural link of the stock market with monetary conditions and from the role that the money supply plays a general indicator of economic expectations. (Homa & Jaffee, 1971). In the past ten years, the importance of the rate of money supply growth to the economy and to the stock market has been increasingly accepted. In stock market analysis, money supply movements are now treated as superior indicators to interpret and provide information about future stock price movements (Rozeff, 1975). Hence, we include money supply as one of the independent variables in our model. We use M2 (broader money) for the money supply variable rather than M1 (narrow money). M2 is defined as an aggregate of currency, demand deposits, other checkable deposits, and travellers’ cheque outstanding, saving deposits and money market deposit account, small time deposits and retail purchase of money market mutual fund (Fisher, 2001). The use of M2 in the model is consistent with the variable used by Hussain.Z & Sohail.N (2009) and Husain & Mahmood (1999) in their studies.

The reason we choose exchange rate act as an important variable to determine stock prices because currency is one of the factors will be considered by foreign investors when they invest a portfolio funds. This is because knowledge about the relationship between currency rates with stock prices will give the investors an accurate estimate of the variability of a portfolio (Dimitrova, 2005). Expensive currency will render the investors to avoid investing in such market with that currency and thus the demand for the particular stock will reduce.

For the case of inflation rate, an increase in inflation will lower down forecasted real earnings growth, that lead to decrease in price-earnings ratio and increase in expected return (Boucher, 2004, 2006). As inflationary shocks may slightly influence the real stock returns, Valcarcel (2012) agreed that there is sensitivity of stock price on inflation within short and medium terms. Campbell (2004) indicated that inflation rate is the main determinant of nominal bond yields which is widely used for years, in which the Fed model also suggest stock yield and inflation are highly correlated. In other words, inflation is also an important factor in explaining our model.

Hypotheses Development

Based on topic of research, we assume that all macroeconomic variables we used will have a direct linear relationship with stock prices.

Economic Function

Stock Price Index = f (money supply, exchange rate, inflation rate)

Economic Model

=β0 + β1 + β2 + β3+

Where,

= Stock price of financial sector in Malaysia at year t

= Money supply (M2) of Malaysia at year t

= Exchange rate (MYR/USD) of Malaysia at year t

= Inflation rate of Malaysia at year t

Expected model

=β0 + β1 - β2 - β3+

Significance of money supply

H0: β1 = 0 (Money supply is not significant in explaining stock price)

H1: β1 ≠ 0 (Money supply is significant in explaining stock price)

Decision rule: We reject H0. if T-statistic of money supply is smaller than negative critical value or T-statistic of money supply is greater than positive critical value, otherwise we do not reject H0..

Significance of exchange rate

H0: β2 = 0 (Exchange rate is not significant in explaining stock price)

H1: β2 ≠ 0 (Exchange rate is significant in explaining stock price)

Decision rule: We reject H0. if T-statistic of exchange rate is smaller than negative critical value or T-statistic of exchange rate is greater than positive critical value, otherwise we do not reject H0..

Significance of inflation

H0: β3 = 0 (Consumer price index is not significant in explaining stock price)

H1: β3 ≠ 0 (Consumer price index is significant in explaining stock price)

Decision rule: We reject H0. if T-statistic of inflation is smaller than negative critical value or T-statistic of inflation is greater than positive critical value, otherwise we do not reject H0..

Overall, we have witnessed the arguments between the researchers regarding on the relationship between the macroeconomic variables and the stock price. The differences between the results are caused by multiple other factors, which include time span, anomalies such as disasters and crises, data obtained, nature of data and research target area. In the following chapters, we will research on how macroeconomic variables affect the stock prices by applying econometrics approach.

CHAPTER 3: METHODOLOGY

3.0 Introduction

In this paper, there are three macroeconomic variables and stock price index used. These variables are in quarterly frequency from Q1 1998 to Q4 2011 which totals up to 56 observations. Detailed elaboration of our methodology is explained in the following parts.

3.1 Research design

As the objectives of the research are about the relationship between stock price and macroeconomic variables, we have focused on quantitative research which involves a series of empirical tests. The dataset for each variable consists of 56 observations which retrieved from Datastream. E-views version 6 software is used to carry those empirical tests to investigate the relationship between dependent variable and independent variables.

3.2 Data collection methods

This paper is mainly focusing on secondary data in times series data. The reasons for using secondary data as our research data are due to its reliability, convenience and time-saving.

3.2.1 Secondary data

All the quarterly stock prices index of financial sector and three independent variables are retrieved from Datastream. The sample period to carry out the investigation spans from year 1998 Q1 to year 2011 Q4. We used FTSE Malaysia Financial stock price index as a proxy of stock price. Financial Trading Stock Exchange (FTSE) Group and Bursa Malaysia joined to provide a set of tradable and investable indices for the Malaysian Market. Thus, it gives a comprehensive range of real-time indices, which covered all eligible companies listed on Bursa Malaysia Main and ACE Markets (FTSE Bursa Malaysia Index Series, n.d). While for macroeconomic variables, it consists of consumer price index (CPI) as the proxy for inflation, money supply (MYR) and exchange rate (MYR/USD).

Variables

Proxy

Explanation

Units

Source of data

Stock price

FTSE

Set of tradable and investable indices

Index

Datastream

Money supply

No proxy used

M2 (Broad money)

MYR in millions

Datastream

Inflation

Consumer Price Index (CPI)

Consumer Price Index in Malaysia

Index number

Datastream

Exchange Rate

No proxy used

Exchange rate in Malaysia

MYR/USD

Datastream

Table 3.1: Source of Data

3.3 Sampling design

3.3.1 Target Population

There are many sectors in Malaysia, while as mentioned we choose financial sector in Malaysia.

3.3.2 Sampling Frame and Sampling Location

Financial sector in Malaysia as our target population which comprises of commercial banks, investments banks, Islamic banks, financial companies, insurance companies and real estate which all will be aggregated into an index.

3.3.3 Sampling Technique

In our study, the sampling technique we applied is judgemental sampling. The population elements are selected based on knowledge and judgment of researchers. A judgmental sample is one in which there is an attempt to draw a representative sample of the population using judgmental selection procedures. The advantages of using this method are it is low cost, convenience and quick. Besides, it is subjective and its value depends entirely on the researches judgment, expertise and creativity (V.E.S College of Arts). In order to avoid crisis after our model, we choose data for post crisis period which is 1998-2011. As we know higher frequency of data provide better results but we use quarterly data since it provides stronger evidence of rather than using monthly data for cointegration (Kasa,1992). Otero & Smith (2000) also found that when using cointegration to investigate relationship between variables, practitioners ought to rely on data collected over a long period of time, rather than on a large number of observations collected over a relatively short period. Since stock price and inflation cannot be measured, proxy variables are used. A proxy variable is a variable that is used to measure an unobservable quantity of interest with a strongly related with each other. For example, FTSE Malaysian Financial provides indices for measuring the stock price of financial sector in Malaysia. Besides, CPI is the most widely used measure of inflation. In order to measure the effectiveness of government economic policy, CPI is used as an indicator (Understanding the Consumer Price Index: Answers to Some Questions, 2004). Thus, FTSE Malaysia Financial stock price index as a proxy of stock price while CPI as proxy of inflation.

Sampling size

We used quarterly data ranging from Q1 1998 to Q4 2011. Therefore there are fourteen (14) years in total, and each year has four (4) quarters. Hence, there are 14 × 4 = 56 observations throughout the research.

Data Processing

A description of data preparation processes such as checking, editing, coding, transcribing, as well as specifying any special or unusual treatments of data before they are analyzed.

Data processing involves four steps. Firstly, the data are obtained from the secondary source. Next, the data will be arranged and edited. The useful data will be analyzed by employing E-views. Lastly, the outcomes and findings are available for interpretation.

We completed our first step by obtaining data from Datastream. The second step would be data transformation and editing. To linearize and smoothen the data curve, we transformed all variables by "logging" it, which means we placed natural logarithm to each set of values. This would also ensure all are under the same units and size as it reduces the seriousness of heterocedasticity because it compresses the scale in which variables are being measured to manageable "chunks". (Kuwornu & Owusu-Nantwi, 2011)

Data analysis

Some empirical tests will be conducted by using E-views version 6 to investigate the relationship of dependent variable and independent variables such Unit Root test, Cointegration test, Vector Error Correction Model (VECM) and Ganger Causality.

Before estimating the model, the stationary of the model variables will be examined by using the augmented Dickey-Fuller (ADF) unit root test. This is essential as the nature of times series data are often found not stationary in many cases. Stationary is defined as mean and variance of the data is zero and constant respectively. Therefore ADF unit root test is employed to check stationary of data. Besides, it can avoid the problem of spurious regression. Stationarity is also important in defining the causality test, especially in time series data set. By fulfilling this assumption, other tests can be run without generating problems (Yahyazadehfar & Babaie, 2012). Here, each variable will be tested in 3 types of level which is at level, at first difference and at second difference along with 2 other options (intercept, intercept and trend)

After running the ADF unit root test, we can know on what degree the data is stationary. If the series are stationary at first difference, then it requires the use of Cointegration test and Vector Error Correction Model for further processing. Hence null hypothesis is non stationary while alternative hypothesis is stationary. (Johansen & Juselius, 1990)

For times series data, lag value is encouraged to include in the model for reducing the autocorrelation problem. This is because last year data is correlated to this year data and the people may not change their behaviour immediately (Shumway & Stoffer, 2011). Before proceeding to cointegration test, the optimal lag value is estimated by VECM. For the quarterly data, the reasonable lag length can range from 1 to 4, 8, 12, etc (Wojcik, 2010). In this paper, we choose quarterly data, so let select 4th lag. The criteria of choosing optimal lag value are based on the lowest Akaike Informtion Criterion (AIC). Many researchers also use AIC as criteria for choosing lag such as Asmy, Rohilina, Hassama, & Fouad (2009) and Adel (2004).

Once the order of integration of each variable has been determined, we employed the Johansen Multivariate cointegration test. This analysis is to estimate whether the time series of these variables display a stationary process in a linear combination. A finding on the cointegration determines the existence of long run relationships between the dependent variable and independent variables. Johansen developed two likelihood ratio tests for testing the number of cointegration vectors (r): the trace test and the maximum Eigenvalue test. If the rank of the matrix is greater than or equal to 1, we reject null hypothesis (no cointegration). However, if that rank is not rejected, we have to proceed to rank 2, and so on till we reach a result of the null hypothesis being rejected. (Johansen & Juselius, 1990)

If there is at least one is cointegrating relationship among the variables, then VECM will be used to estimate the relationship of variables. This is due to Johansen VECM is a full information maximum likelihood estimation model which allows in testing the whole system in one step (Maysami, Howe, & Hamzah, 2004). Although OLS regressions may produce highly significant parameters, it raises the problem of autocorrelation in the regression disgnostics in time series data. Time series data has high tendencies in displaying autocorrelation problems (Hoxha, 2010). Therefore, the VECM models are frequently applied in examining models with more than one endogenous variable and in time series. (Hoxha, 2010)

Further, if the series are cointegrated the Granger causality test must be estimated in a Vector Correction Model (Rossini & Kupke, 2012). Clive Granger proposed Granger Causality technique in order to determine causality between two time series and whether one time series is useful in forecasting another. It is used to estimate causal relationship and exist of unidirectional or bidirectional between dependent variable and independent variables. (Harasheh & Abu-Libdeh, NA). Many previous researches used Granger Causality test to analyse the relationship between stock prices and macroeconomic variables (Granger, Huang, & Yang, 1998).

Lastly, to ensure the Vector Error Correction Model (VECM) is free from econometric problems, there are some tests are needed to be conducted such as the heteroscedasticity test, autocorrelation test and normality test.

Autocorrelation is defined as the error terms are correlated to each other. This leads to biased, inefficient and inconsistent result. This econometric problem can be tested by using Breusch Lagrange multiplier (LM) test.

If the variance of the errors is not constant, it means the problem of heteroscedasticity exists in the model. This econometric problem can be tested by using White-test.

Normality Test and Jarque-Bera (JB) test statistic will be used to test the normality of a model. JB test uses the property of a normally distributed random variable that the entire distribution is characterized by first two moments-mean and median. While the standardized third and fourth moments of a distribution are skewness (S) and kurtosis (K). For a normally distributed variables, S=0 and K=3

If three assumptions are violated, the estimator will be inefficient and no longer best linear unbiased estimator (BLUE). (Brooks, 2008).

Conclusion

By employing the methods above, a proper data analysis could be done for our quantitative research model, equipped with necessary quarterly data from the stock prices index of financial sector in Malaysia from the year 1998 to 2011 as well as the macroeconomic variables (money supply, real exchange rate and inflation rate) data. The following chapter would present our findings on the analysis and empirical tests.

CHAPTER 4 : DATA ANALYSIS

4.0 Introduction

After our data collection and transformation is completed, we proceed to analyse our data. We used E-Views version 6 program to estimate the model in order to investigate the relationship between stock price financial sector that are listed on Bursa Malaysia, and three macroeconomic variables. In total, there are six different tests were conducted by running the E-Views version 6 program, namely Unit Root Test, Cointegration Test, Vector Error Correction Model (VECM), Granger-causality Test and Diagnostic Tests (heteroscedasticity test, autocorrelation test and normality test). The outcome of each test would be further explained below.

Unit Root Test

Augmented Dickey-Fuller (ADF) was tested to detect the unit root on stock price, money supply, real exchange rate and inflation. We conducted two different types of stationary process: first, constant; second, constant with linear trend. The null hypothesis is the variable is not stationary. The acceptance or rejection of the hypothesis is based on MacKinnon (1996) one-sided p-values to significant level of 5% (0.05). It showed that money supply, exchange rate and consumer price index are not stationary at level, except for stock price as it is stationary at constant with linear trend. All the variables are stationary at first difference for two different type of stationary process, in which it indicates that both independent and dependent variables are cointegrated at order one or I(1). The ADF outcome is shown in Table 1.

Table 1 Augmented Dickey-Fuller Unit Root Test

Variable

Level

Constant

Constant with linear trend

LSPI

-1.936819

-6.132152*

LMS

1.188423

-1.971479

LEX

-0.977832

-3.383193

LCPI

0.125654

-2.397620

Variable

First Difference

Constant

Constant with linear trend

LSPI

-4.742890*

-4.672993*

LMS

-5.987060*

-5.994178*

LEX

-3.797061*

-4.017287*

LCPI

-6.131257*

-6.130545*

The values refer to Augmented Dickey-Fuller test statistic.

* indicates significance at 5%.

Cointegration

Before proceeding to cointegration test, we need to minimize the error of the empirical result, that is, by adding lag variables. The lag length selection was done based upon the lowest value of Akaike Information Criteria (AIC). As shown in Table 2, AIC has the lowest value in lag order of three. Thus, we inserted three lag variables into our models.

Table 2 Akaike Information Criteria (AIC)

Lag

AIC

1

-1.227030

2

-1.819604

3

-1.973102*

4

-1.553969

* indicates the lowest value

Trace test and Maximum Eigenvalue test were used by applying lag interval of three that is obtained from the AIC value previously. The null hypothesis for the tests is there is no cointegration. As mentioned earlier, the acceptance or rejection of the hypothesis is based on MacKinnon-Haug-Michelis (1999) p-values to significant level of 5% (0.05). The outcome shows that there are two cointegrating vectors for both tests, as shown in Table 3. Thus, there is a long run relationship between dependent variable and independent variables.

Null Hypothesis

Alternative Hypothesis

Trace

Max-Eigen

r=0

r≥1

87.31846*

46.40986*

r ≤ 1

r≥2

40.90860*

31.34503*

r≤2

r≥3

9.563567

9.538505

r≤3

r≥4

0.025061

0.025061

Table 3 Trace and Maximum Eigenvalue Tests Statistics

* indicates rejection of the hypothesis at 5%

Vector Error Correction Model

As our model has two cointegrating vectors (noting that there is long run relationship existing between the variables), Vector Error Correction Model was used to determine the short term relationship of stock price index and all the macroeconomic variables, and which macroeconomic variable is important in explaining stock price index. The equation of Vector Error Correction Estimates for stock price index is shown as below:-

LSPI = 3.957144 + 0.846351LMS – 2.403233LEX – 2.212137LCPI + εt

(-1.54203) (3.14034) (0.76182)

( ) indicate t-statistics

From the equation above, all the signs that we obtained from Vector Error Correction Estimates are consistent with expected sign of money supply, exchange rate and consumer price index in Chapter 2. Money supply has insignificant (-1.54203>-2) and positive relationship with stock price index. For 1% increase in money supply, on average, stock price index will increase by 0.846351%, and vice versa. As for exchange rate, it is significant (3.14034>2) in explaining stock price index. There is negative relationship between both stock price index and exchange rate. If exchange rate is depreciated by 1%, on average, stock price index will increase by 2.403233%, and vice versa. As for consumer price index, it has insignificant (0.76182<2) and negative relationship with stock price index. Every 1% decrease in consumer price index, on average, stock price index will increase by 2.212137%, and vice versa.

Granger Causality Test

Table 4.1: Granger Causality Tests

Dependent

variables

Independent variables (F Statistic)

ΔLSPI

ΔLMS

ΔLEX

ΔLCPI

ΔLSPI

-

19.36835*

(0.002)

0.708976

(0.8711)

16.35086*

(0.0010)

ΔLMS

0.456398

(0.9284)

-

-

-

ΔLEX

1.251729

(0.7406)

-

-

-

ΔLCPI

4.608507

(0.2028)

-

-

-

* indicates significance at 5%.

Granger causality test is used to predict the short run relationship such as unidirectional or bidirectional exists between variables. The general null hypothesis for Granger causality test is that there is no Granger cause relationship between variable 1 and variable 2, where the variables noted are the ones we are examining with. The decision rule to reject the null hypothesis would be based on the comparison of p-value obtained with the significant level of 5%.

Money Supply (MS)

From Table 4, the p-value (0.002) is smaller 5% significant level. Thus, the result shows that financial sector stock price index is Granger caused by money supply.

Exchange Rate (EX)

From Table 4, the p-value (0.8711) is greater than 5% significant level. Thus, the result shows that financial sector stock price index is not Granger caused by exchange rate.

Consumer Price Index (CPI)

From Table 4, the p-value (0.0010) is smaller than 5% significant level. Thus, the result shows that financial sector stock price index is Granger caused by consumer price index.

Table 4.1 presented the causal relations of each of the independent variables to the dependent variable. Based on the hypothesis above, we can conclude there is no bidirectional or unidirectional relationship of stock price index to the examined macroeconomic variables. However there are unidirectional relationships of money supply to stock price index and consumer price index to stock price index but the result shows no causal relationship between stock price and exchange rate.

Diagram 4.2: Analysis between Dependent Variable and Independent Variables for Granger Causality Relationship

ΔLMS

ΔLEX

ΔLSPI

ΔLCPI

Indicator:

One way causal relationship

No causal relationship

VEC Residual Diagnostic Test

In order to perform an accurate and reliable result, we have performed a VECM diagnostic checking for auto correlation test, heteroscadaticity test, and normality test after unit root test. The general null hypothesis for diagnostic checking is there is no auto correlation problem, no heteroscadaticity problem, and the error term is normally distributed. The decision rule to reject the null hypothesis is based on the p-value to significant level of 5%. Thus, we reject the null hypothesis if p-value of the tests is less than 5%.

Autocorrelation

Table 5.1: VEC Residual Serial Correlation LM Tests

Lags

LM-Stat

Prob

3

 13.90652

 0.6057

Form table 5.1, the p-value (0.6057) is greater than 5% significant level. Thus, we can conclude that the null hypothesis is not rejected and it proven there is no auto correlation problem exists.

Heteroscadasticity

Table 5.2: VEC Residual Heteroscadaticity Test

Chi-sq

df

Prob.

 220.9777

260

 0.9621

From table 5.2, the p-value (0.9621) is greater than 5% significant level. Thus, we can conclude that the null hypothesis is not rejected and it proven there is no heteroscadaticity problem exists.

Normality Test

Table 5.3: VEC Residual Normality Test

Component

Jarque-Bera

df

Prob.

Joint

 10.33475

8

 0.2423

From table 5.3, the p-value (0.2423) is greater than 5% significant level. Thus, we can conclude that the null hypothesis is not rejected and it proven the error term is normally distributed.

Conclusion

In this chapter, all the empirical results have presented clearly in table form or figure form. Besides, clear and precise explanation has been written for all the empirical results. Further, the summary of the discussion, conclusion, and policies implication will be presented in next chapter

References Chapter 1-3

Abraham, T. W. (n.d.). Stock Market Reaction to Selected Macroeconomic Variables in the Nigerian Economy. CBN Journal of Applied Statistic , 2 (1), 61-70.

Adel, A.-S. (2004). The Dynamic Relationship Between Macroeconomic Factors and The Jordanian Stock Market. International Journal of Applied Econometrics and Quantitative Studies , 1 (1), 97-114.

Ahmad, H. (2005). The Effects of Banking Sector and Stock Market Development on the Malaysian Economic Growth: An Empirical Investigation. Thesis , 1-24.

Ahmad, Z., Ahmad, Z., Khan, M. S., & Javaid, U. (2012). Capturing the stock price movements at Karachi Stock exchange: Are macroeconomic variables relevant? African Journal of Business Management , 6 (8), 3026-3034.

Ahmad, Z., Ahmad, Z., Khan, M. S., & Javaid, U. (2012). Capturing the stock price movements at Karachi Stock exchange: Are macroeconomic variables relevant? African Journal of Business Management , 6 (8), 3026-3034.

Alper, C. A. (2007). Estimating the Effects of Interest Rates on Share Prices Using Multi-scale Causality Test in Emerging Markets: Evidence from Turkey. Munich Personal RePEc Archive .

Arango. (n.d.). Returns and interest rate: A nonlinear relationship in the Bogotá stock market . Banco de la República , 15.

Asmy, M., Rohilina, W., Hassama, A., & Fouad, M. (2009). Effects of Macroeconomic Variables on Stock Prices in Malaysia: An Approach of Error Correction Model. Munich Personal RePEc Archieve .

ASPREM, M. (1989). STOCK PRICES, ASSET PORTFOLIOS AND MACROECONOMIC VARIABLES IN TEN EUROPEAN COUNTRIES . Journal of Banking and Finance , 589-612.

Aydemir, O., & Demirhan, E. (2009). The Relationship between Stock Prices and Exchange Rates Evidence from Turkey. Journal of Finance and Economics (23), 207-215.

Baharom, A., Royfaizal, R., & Habibullah, M. S. (2008). Pre and Post Crisis Analysis of Stock Price and Exchange Rate: Evidence from Malaysia. International Applied Economics and Management Letters , 1 (1), 33-36.

Bernanke, B. S., & Kuttner, K. N. (2004). What Explains the Stock Market’s Reaction to Federal Reserve Policy? Working Paper .

Birz, G., & Jr, J. R. (2011). The effect of macroeconomic news on stock returns: New evidence from newspaper coverage. Journal of Banking & Finance (35), 2791–2800.

BOSTON, T. F. (1971, J