Indonesian banking sector thesis methodology
Published:
This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
CHAPTER III
METHODOLOGY
3.1 Research Design
According to Burns and Grove (2003:195), a research design is blueprint for conducting a study that maximizes control over factors that may interfere with the validity of the findings. It is the end result of a series of decision made by the researcher concerning how the study will be conducted and closely associated with the framework of the study that guides the plan for implementing the study. While according to Polit et al (2001:167), a research design is the researcher’s overall for answering the research question or testing the research hypothesis. The research design for most quantitative studies are highly structured while in qualitative studies are more flexible.
This thesis used quantitative method to identified, analysed, and describe the influence of intellectual capital towards the financial performance in the Indonesia banking sector. According to Aliaga and Gunderson (2010), quantitative research is explaining phenomena by collecting numerical data that are analysed using mathematically based methods (in particular statistics).
According to Burns and Grove (2003:27-30), Quantitative research has the following characteristics, such as:
There is a single reality that can be defined by careful measurement.
- It is usually concise.
- It describes, examines relationships, and determines causality among variables, where possible.
- Statistical analysis is conducted to reduce and organise data, determine significant relationships and identify differences and/or similarities within and between different categories of data.
- The sample should be representative of a large population.
- Reliability and validity of the instruments are crucial.
- Comprehensive data collected by employing different methods and/or instruments should result in a complete description of the variable or the population studied.
- It provides an accurate account of characteristics of particular individuals, situations, or groups.
3.2 Sampling Design
The population in this thesis is the banking sector companies listed in Indonesia Stock Exchange. This study uses financial statement data for the last five years from 2008-2012 to provide an overview of the company's performance. Sampling method in this thesis is purposive sampling method with multiple criteria. These criteria are banks that listing in Indonesia Stock Exchange and has published financial reports in 2008 until 2012 as well as did not experience a loss in the reporting year. Criteria of not experiencing lose in this thesis is specified with the aim to keep the company growth measurements being positive. Based on the predefined criteria, samples used were 23 banks and with the method of merging the data, thus resulting in 115 observational data. Sample selection criteria are presented in Table 3.1 as follow:
Table 3.1: Determination of Sample Size
Criteria |
Number of Company |
Number of Observational Data |
Banks listed in Indonesia Stock Exchange in 2008, 2009, 2010, 2011, 2012 |
26 |
26 x 5 = 130 |
Banks that suffered losses in 2008, 2009, 2010, 2011, 2012 |
3 |
3 x 5 = 15 |
The number of sample used in this thesis |
23 |
23 x 5 = 115 |
Source: ICMD and IDX, 2013
3.3 Research Instruments
3.3.1 Data Collection Method
The data collection methods used in this study is content analysis, a method of collecting research data through observation and analysis techniques to the content or message of a document. The goal of content analysis is to identify the characteristics or specific information contained in a document to produce an objective and systematic description (Indriantoro, 2002).
Content analysis is done by reading the annual report of each banks’ sample and code the information contained therein. The steps in conducting content analysis according Bozzolan et al. (in Purnamosidhi, 2003) include (1) selecting a framework that is used to classify the information, (2) define the recording units, (3) coding, (4) assess the level of reliability is achieved.
3.3.2 Sources of Data
The data used in this study is secondary data. The secondary data the research is obtained by researcher indirectly through an intermediary medium (obtained and recorded by others) in the form of evidence, historical records or reports that have been arranged in the archive (documentary data) published and unpublished (Indriantoro and Supomo, 2002). The data obtained for this thesis is obtained from the financial statements of banking sector listed on the Stock Exchange in 2008 -2012. Other sources of data is obtained from the Indonesian Capital Market Directory.
3.4 Operational Definition
3.4.1 Independent Variable
The independent variable in this study is the intellectual capital. According to Williams, the intellectual capital is information and knowledge that is applied in the work to create value (Purnomosidhi 2006). Currently, an effort to provide an assessment of intellectual capital is essential. Pulic (1998) proposed a Value Added Intellectual Coefficient (Value Added Intellectual Capital/ VAIC) to provide information about the value creation efficiency of tangible and intangible assets in the company.
VAIC is an analytical procedure that is designed to enable the management, shareholders and other relevant stakeholders to effectively monitor and evaluate the efficiency of the total value added to the company's resources and the individual components of the primary resource. Value added (VA) is the difference between sales (OUT) and input (IN). The formula to calculate the VA is:
VA = OUT – IN
Where:
OUT = the revenue and comprise all the products and services sold on the market.
IN = all the expenses incurred in earning the revenueexcept manpower costs
VAIC measure the efficiency of the three types of company’s input: human capital, structural capital and physical capital and financial,
3.4.1.1 HCE
Human Capital (HC) refers to the collective value of the intellectual capital of the company, namely competence, knowledge, and skills (Pulic, 1998; Firer and Williams, 2003), which is measured by the Human Capital Efficiency (HCE). HCE is an indicator of the efficiency of the value-added (value added/ VA) of human capital. The formula to calculate the HCE is:
HCE =
Where:
HCE = Human Capital Efficiency
VA = Value Added
HC = Human Capital
= Total salary and allowance cost
Salary is a form of remuneration or award given to an employee on a regular basis for services and their work. Allowances are elements of remuneration are given in rupiah directly to individual employees and can be known with certainty. Allowance is given to employees in order to induce/ improve morale for employees.
3.4.1.2 SCE
Structural Capital (SC) can be defined as competitive intelligence, formulas, information systems, patents, policies, processes, and so on, as the result of the company's product or system that has been created from time to time (Pulic, 1998; Firer and Williams, 2003), which is measured by Structural Capital Efficiency (SCE). SCE is an indicator of the efficiency of the value-added (value Added / VA) of structural capital. The formula for calculating the SCE is:
SCE =
Where:
SC = VA – HC
SCE = Structural Capital Efficiency
SC = Company’s Structural Capital
VA = Value Added
HC = Human Capital
3.4.1.3 CEE
Capital Employed (CE) is defined as the total capital used in fixed and current assets of a company (Pulic, 1998; Firer and Williams, 2003), which is measured by Capital Employed Efficiency (CEE). CEE is an indicator of the efficiency of the value-added (Value Added / VA) of capital employed. The formula for calculating the CEE is:
CEE =
Where:
CEE = Capital Employed Efficienncy
VA = Value Added
CA = Capital Employed
= Total Asset – Intangible Asset
3.4.1.4 VAIC
The value of VAIC can be obtained by adding the three components namely HCE, SCE and CEE. VAIC indicate an organization's intellectual abilities. VAIC can also be considered as BPI (Business Performance Indicator). The formula for calculating VAIC namely:
VAIC = HCE + SCE + CEE
3.4.1 Dependent Variable
3.4.1.1 Return on Asset (ROA)
ROA describe the business benefits and efficiency in the utilization of the company's total assets (Chen et al, 2005). The formula for calculating ROA is as follow:
ROA =
Where:
Net Income includes net interest income and non interest income.
3.4.1.2 Return on Equity (ROE)
Another dependent variable in this thesis is the performance of the company which is proxied by return on equity (ROE). ROE is a financial indicator that describes the company's ability to generate a return on total assets of the company. The formula for calculating ROE is as follow:
ROE =
3.5 Data Analysis Procedure
3.5.1 Descriptive Statistic
Descriptive statistics provide a picture or description of the thesis variables which seen from the average value (mean), minimum, maximum and standard deviation (Ghozali, 2006). Maximum and minimum show the largest and smallest values. Overview of these data yield clear information and thus the data is easily understood. In this thesis, by seeing an overview of the data available, it will give clear information about the influence of intellectual capital on the financial performance of the company.
3.5.2 Classical Linear Regression Model Assumption Test
By using Original Least Square (OLS) equation in calculating regression analysis, there are several assumptions that must be met in order for the regression equation to be valid for use in research. These assumptions are called the classical linear regression model assumptions. It includes the normality test, multicollinearity test, autocorrelation test, and heteroscedasticity test.
3.5.2.1 Normality Test
Normality test aims to test whether the regression model, the dependent variable, independent variables, or both have a normal distribution or not. Regression model is considered good when it has a normal data distribution or dissemination of statistical data on the diagonal axis of the normal distribution graph.
The way to see whether the data are normally distributed or not is by using the P-plot graphs and statistical tests of non-parametric Kolmogorov-Smirnov (KS). Normally distributed data is when the results of the P-plot is near the diagonal line and the results of the Kolmogorov-Smirnov test showed significant values â€‹â€‹above 0.05 (Ghozali, 2006).
3.5.2.2 Multicollinearity Test
Multicollinearity test is intended to test whether there is a correlation between the independent variables in the regression model. The test for multicollinearity is done by looking at Tolerance and Variance Inflation Factor (VIF) value. Both of these measurements indicate which of each independent variable is explained by other independent variables. In simple terms each of the independent variables is the dependent variable and regressed on the other independent variables. Tolerance measures the variability of the selected independent variables that are not explained by the other independent variables (Ghozali, 2006). Therefore, a low Tolerance value equal to the high value of VIF (since VIF = 1/Tolerance). Cutoff value that is commonly used to indicate the presence of multicollinearity are tolerance values of â€‹â€‹<0.10 or equal to the value of VIF> 10 (Ghozali, 2006).
3.5.2.3 Autocorrelation Test
The autocorrelation test is used to test whether there are correlation between the residuals in the linear regression model (Ghozali, 2009). To detect whether there is autocorrelation is by using the Durbin Watson test. The decision whether there is autocorrelation according to Imam Ghozali (2009) are as follow:
Table 3.2: Durbin Watson Autocorrelation Decision
Hypothesis |
Decision |
If |
No Positive Autocorrelation |
Reject |
0 < d < |
No Positive Autocorrelation |
No Desicion |
< d < |
No Negative Autocorrelation |
Reject |
4- < d < 4 |
No Positive Autocorrelation |
No Desicion |
4- < d < 4- |
No Autocorrelation, both positive nor negative |
Do Not reject |
< d < 4- |
Source: Tarigan (2011)
Autocorrelation test aims to test whether in the linear regression model there is a correlation between the error in period t with error in period t-1 or earlier periods. Another way to find out whether or not there is autocorrelation is by using the Run Test. Run Test as part of the non-parametric statistics can be used to test whether there is a correlation between high residual. If there is no relationship between the residual correlations it is said that the residuals are random. Run test is used to see if the data residuals occur randomly or not (systematically). Autocorrelation is not significant if the probability is greater than a = 0.05 (Ghozali, 2006).
3.5.2.4 Heteroscedasticity Test
Heteroscedasticity test aims to test whether in the regression model occurs inequality variance of the residual in one observation to another observation. If variance of residual in one observation to other observations remain, then it is called homoscedasticity and if different it is called heteroscedasticity.
The way to determine whether there is heteroscedasticity or not is to look at the graph plots between the predicted values of the dependent variable which is ZPRED with the residual of SRESID. Heteroscedasticity does not happen if there is no clear pattern and the points spread above and below the number 0 on the Y axis (Ghozali, 2006)
3.5.3 Linear Regression Analysis
This study will test eight hypotheses that have been developed. To test the hypotheses, four models of regression in this thesis are as follow:
Model I:
Model II:
Model III:
Model IV:
Regression analysis is used to measure the strength of relationship between two or more variables. In addition, regression analysis is conducted to demonstrate the positive or negative direction of the relationship between the dependent variable with the independent variable (Ghozali, 2009).
3.5.4 The coefficient of determination (R^{2})
The coefficient of determination (R^{2}) measures the extent to which the ability of the model to explain the variation in the dependent variable (Ghozali, 2009). From here will be known how much the dependent variable explained by the independent variables, while the rest is explained by other causes outside the model.
The higher the coefficient of determination, the better the ability of the independent variables in explaining the dependent variable. There are two types of coefficient of determination is the common coefficient of determination and the adjusted coefficient of determination/ Adjusted R Square (Purbayu and Ashari, 2005).
The coefficient of determination is between zero and one. A small value means the ability of the independent variables in explaining variation in the dependent variable is very limited. Value close to one means that the independent variables provide almost all the information needed to predict the variation in the dependent variable.
3.5.5 Significance F-Test
F-test statistics basically show whether all the independent variables which are included in the model have jointly influence the dependent variable (Ghozali, 2006). The way to find out is by comparing the calculated F value with the value of F table. If the calculated F value is greater than the value of F table, then the alternative hypothesis is accepted meaning that all the independent variables together and significantly affect the dependent variable. It can also be seen by a probability. If the probability (significance) less than 0.05 (a) the independent variables together (simultaneous) effect on the dependent variable.
3.5.6 Significance T-test
Statistical T-test basically shows how far the influence of the independent variables in explaining the variation of individual dependent variables. The way to find out is to compare the t value with the value of t table. If the t value is greater than the value of t table, it means the t value is significant meaning the alternative hypothesis is accepted and that individual independent variables affect the dependent variable. In addition, it could also be done by looking at the p-value of each variable. The hypothesis is accepted if the p-value <5% (Ghozali, 2007).