In the research design, stock returns figures and the respective macroeconomic variables that are perceived to have logical and economic relationships with stock returns will be employed. In order to proxy for stock returns, Straits Times Index (STI) will be used. The data concerning STI will be obtained from Yahoo Finance (http://sg.finance.yahoo.com/). In the database, the data for daily, monthly, and quarterly STI is available, starting from 1 January 1988. For the purpose of this dissertation, monthly data will be used. Thus, monthly closing price of STI of the respective month historically will be obtained from the website. This series of data will be used to proxy stock returns in Singapore.
From another end, the data of macroeconomic variables, which include the interest rates, inflation rate, money supply, and Industrial Production Index (IPI) for Singapore, can be purchased from Singapore Department of Statistics. The address and contacts of Singapore Department of Statistics is as follow: Singapore Department of Statistics, Statistical Information Services, 100 High Street #05-01, The Treasury, Singapore 179434 [Fax No. : 65-63327689]. The details on each of these macroeconomic variables will be further discussed in the next section.
In this section, the various variables used in this study will be discussed and explained. As discussed before, stock returns in Singapore will be represented by closing price of STI at the end of each month. As the macroeconomic variables, are only available on monthly basis (but not daily), monthly stock returns figures will be used in this analysis. In order to calculate the growth rate of STI from a month to the next, the following formula will be used.
To explain, Change_STI is the symbols used to represent the growth rate of STI from a month to the next. As shown in the formula above, it is calculated by the rate of change of STI from a particular month to the next. The rate of change of STI is employed int his analysis, as it is the objective of this study to understand how the change in macroeconomic variables relate or associate with the rate of change in STI.
From another perspective, there are many variables that can be used to represent the different categories of macroeconomic variables in Singapore. As discussed before, the different categories of macroeconomic variables that will be investigated include the following: interest rates, inflation rate, money supply, and Industrial Production Index (IPI) in Singapore. For interest rates, there are actually different types of interest rates available from the database provided by Singapore Department of Statistics. All of the interest rates figures are available monthly. In this dissertation, four types of interest rates will be used to investigate the relationships of stock returns to interest rates. The four interest rates are: (a) 1 month interbank rate (i.e., 1m_Interbank), (b) 3 month interbank rate (i.e., 3m_Interbank), (c) yield of 1 year Treasury bill (i.e., 1y_Treasury), and (d) yield of 10 years Treasury Bonds (i.e., 10y_Treasury). For ease of reference, the symbols used to represent each of these interest rates are shown in bracket as above. However, the level of these interest rates will not be used in the study of interest rates relationships to stock returns. The rate of change of the interest rate, however, will be employed. The symbols used to represent the rate of change for each of the respective interest rates are shown in Table 3.1 below.
In order to proxy for real economic activities, Industrial Production Index (IPI) of Singapore will be employed. Similarly, the rate of change for IPI will be used as the input in the subsequent analysis. The rate of change of IPI is represented by the following symbol: change_IPI. Then, the inflation rate of Singapore, however, will be represented by Consumer Price Index (CPI) in Singapore. Again, the rate of change for CPI will be used as the input in the subsequent analysis. The rate of change of CPI is represented by the following symbol: change_CPI.
Then, on the other hand, money supply used in this dissertation include the three definition of money supply often discussed in finance academic textbook, namely M1, M2 and M3. Similar to IPI and CPI, the rate of change for money supply will be used as the input in the subsequent analysis. The rate of change of money supply, namely, M1, M2 and M3 employed are represented by the following symbols respectively: change_M1, change_M2 and change_M3.
In order to study the relationships between stock returns, as represented by rate of change in STI to the rate of change of the other macroeconomic variables, three statistical concepts or tools will be employed, namely, Pearson correlation coefficient, regression analysis (i.e., represented via ANOA table) and scatter plot (for graphical analysis of the relationships between two variables).
Firstly, correlation between the stock index returns to the various economic indicators, with data related to Singapore will be used. Ability to discover statistical significant correlation coefficients will indicate that economic indicators are perhaps useful and valuable in making prediction about future stock market returns, or vice versa. The existence of statistically significant correlation coefficient between two variables suggest that there is association between the two variables, and that can lead us to further investigate how the particular two variables related to each other in greater depth.
Then, the two variables, namely, the stock returns as well as the rate of change of the respective macroeconomic variables, will be analyzed through regression analysis. ANOVA table (i.e., analysis of variance) for the regressions will be simulated. From the ANOVA table, the statistically significance of the regression can be analyzed. Then, in order to provide graphical representation for reader to further understand if how the two variables associate with each other, scatter plot will be generated. Scatter plot will assist the researcher to identify if the regression is indeed affected by certain outliers, and hence, make subjective judgment if the statistical numbers generated from the regression is reliable or dependable.
It is acknowledged that the study carried out in this dissertation is nonetheless, a simple and preliminary one. There are several limitations associated with the nature of this research design, and deserve attentions from readers.
Firstly, the statistical method of correlation coefficient and linear regression assume existence of linear relationships between stock returns to the macroeconomic variables being investigated. However, the real relationships between macroeconomic variables to stock returns may be more dynamic or complex than people can ever imagine. In fact, linear relationships between stock returns and macroeconomic variables, is simply not rational or logical – as the different macroeconomic variables and stock returns may influence or affect each other from time to time. As such, to assume linear relationships may be simplisistic, and may cause inaccurate interpretation of the output generated from the analysis or findings to be presented in the next chapter.
Besides, it is also assumed that throughout all the period being study, the relationships between macroeconomic variables to stock returns remain constant. In other words, it is ignored explicitly that the relationships between macroeconomic variables and stock returns may change according to different time period or regime. In other words, structural change and its influences towards the relationships to be observed are simply ignored. However, as discussed previously in the literature review in Chapter 2, it is understood that certain scholars or researchers had been identifying empirical evidences supporting the fact that relationships between macroeconomic variables to stock returns may differ in different phase of the business cycle. Indeed, under different market sentiment, it can be reasonably expected that investors may have different expectation on the release or announcement of macroeconomic indicators, or simply being affected differently by the changes of economic climates at that particular point of time. As such, it is important to keep in mind that the findings from this study may be affected by the structural change or the impacts from stages of business cycles relevant to Singapore.
Then, it is also crucial to understand that the discovery of statistical significant relationships between stock returns to macroeconomic variables may not indicate investors will definitely have the opportunities to exploit superior or excess returns from the financial markets in Singapore. Firstly, it may be true that such relationships may no longer stay constant in the future. Secondly, correlations or regression analysis is probably best used to analyzed associations between the two variables, but it is unsure if stock returns lead macroeconomic variables or otherwise. The causation between stock returns to macroeconomic variables is not investigated. To do that, other statistical methods, such as Granger causality analysis will be required. Thus, it is crucial that reader to keep in mind that results findings from this study should be treated as preliminary study that guide investors for further study or researches to be conducted to more confidently identify how the macroeconomic indicators or news announcements can be employed to yield excess returns from the challenging stock market.