CHAPTER 3 RESEARCH METHODOLOGY

Transcription

1 CHAPTER 3 RESEARCH METHODOLOGY 3.1 Aim and Framework The aim of this research is to investigate the influence of some key factors on Return on Assets (ROA) of non-financial companies listed in Kompas 100 index at Indonesia Stock Exchange (IDX). The author uses the framework as follows: Research question Literature review Hypothesis development Model building Model 1 Working capital management Model 2 CCC elements Data collection: Annual report Descriptive statistic analysis Test assumption analysis Multiple regression analysis Result 20

2 21 Figure 3.1 Research framework 3.2 Variables The independent variables are Current Ratio (CR), Cash Conversion Cycle (CCC), Net Working Capital Level (NWCL), daily Working Capital (DWC), ARP, APP, and IP. Dependent variable is Return on Assets (ROA). Each of the independent variables effects towards ROA will be investigated. Table 3.1 Variables and calculation method Variables type Variables name Measurement Dependent ROA Net profit/total assets Independent - ARP - APP - IP - CR - NWCL - CCC - DWC 365/(sales/accounts receivable) 365/(COGS/accounts payable) 365/(COGS/inventories) Current assets/current liabilities (current asset-current liabilities)/total assets (ARP+IP) APP (receivables+inventories)-payables/daily sales 3.3 Model Building The author developed two models on this research to explain the return of total assets through the indicators shown in table 3.1. These models are: Model 1 Model 2 Where RTA : RTA i = β 0 + β 1 CCC + β 2 DWC + β 3 CR + β 4 NWCL + ε : RTA i = β 0 + β 1 ARP + β 2 IP + β 3 APP +β 4 CR + β5 NWCL + ε = Return on Assets CCC = Cash Conversion Cycle DWC = Daily Working Capital

3 22 CR = Current ratio NWCL = Net Working Capital Level ARP IP APP = Accounts Receivable Period = Inventory Period = Accounts Payable Period In model 1, the author uses four independent variables; CCC, DWC, NWCL, and CR. CCC is used to measure the length of time funds are tied up in working capital, DWC is used to measure the company s need for capital, NWCL is used to measures net working capital proportion to the total assets, and CR is used to measure the liquidity of the firms. In model 2, the author tried to deal with the elements (accounts receivable, inventory, and accounts payable) of the findings about cycle time in model 1. ARP is used to measure day s sales outstanding. APP is used to measure day s payable outstanding and IP is used to measure inventory holding period. 3.4 Hypothesis The purpose of this research is to prove that there is a relationship between independent variables against dependent variable. The hypotheses that will be tested in this research are as follows: Hypothesis 1:

4 23 H 0 : No relationship between the dependent variable (ROA) and the independent variables (CR, CCC, NWCL, DWC). H 1 : There is relationship between the dependent variable (ROA) and the independent variables (CR, CCC, NWCL, DWC). Hypothesis 2: H 0 : No relationship between the dependent variable (ROA) and the independent variables (ARP, IP, APP). H 1 : There is relationship between the dependent variable (ROA) and the independent variables (ARP, IP, APP). 3.5 Population and Sample This study focuses on non-financial companies listed in Kompas 100 index at Indonesia Stock Exchange (IDX) as of February 2010 for hypothesis testing. After selection, the author draws back 15 years to get the financial report data from The author chooses this period since it reflects the firms condition pre and post crisis period. After the selection, there are 31 eligible non-financial companies that meet the criteria. Table 3.2 Sample companies listed in Kompas 100 index No Code Company Name 1 ADES Akasha Wira International Tbk 2 ADMG Polychem Indonesia Tbk 3 ASGR Astra Graphia Tbk 4 ASII Astra International Tbk 5 BLTA Berlian Laju Tanker Tbk 6 BNBR Bakrie and Brothers Tbk 7 BRPT Barito Pacific Tbk 8 BUMI Bumi Resources Tbk

5 24 9 CPIN Charoen Pokphand Indonesia Tbk 10 CTRA Ciputra Development Tbk 11 GGRM Gudang Garam Tbk 12 GJTL Gajah Tunggal Tbk 13 INCO International Nickel Indonesia Tbk 14 INKP Indah Kiat Pulp and Paper Tbk 15 INTP Indocement Tunggal Prakarsa Tbk 16 JPFA Japfa Comfeed Indonesia Tbk 17 JPRS Jaya Pari Steel Tbk 18 KLBF Kalbe Farma Tbk 19 MPPA Matahari Putra Prima Tbk 20 MTDL Metrodata Electronics Tbk 21 MYOR Mayora Indah Tbk 22 RMBA Bentoel International Inv Tbk 23 SMCB Holcim Indonesia Tbk 24 SMGR Semen Gresik Tbk 25 SOBI Sorini Agro Asia Corporindo Tbk 26 SULI Sumalindo Lestari Jaya Tbk 27 TKIM Pabrik Kertas Tjiwi Kimia Tbk 28 TSPC Tempo Scan Pacific Tbk 29 UNSP Bakrie Sumatra Plantations Tbk 30 UNTR United Tractors Tbk 31 UNVR Unilever Indonesia Tbk 3.6 Data Collection Data used in this study is secondary data. The data needed are CR, CCC, NWCL, DWC, ARP, APP, IP and ROA from financial statement of the nonfinancial companies listed in Kompas 100 index at IDX as of February Data Analysis Method This study focuses on non-financial companies listed in Kompas 100 index at IDX. Descriptive statistics, correlation analysis including classical assumption test and multiple regression analysis are used to generate findings.

6 Descriptive Statistic Descriptive statistic describes minimum, maximum, average and standard deviation of the samples so that the reader can have a thorough understanding towards the variables and the sample characteristic Correlation Analysis Correlation analysis is done to trace the mutual influences of variables one another. This research uses Pearson Coefficient of Correlation Multiple Regression Analysis Multiple regression analysis is a statistical technique to predict the variance in the dependent variable by regressing the independent variables against the dependent variable. The regression will be calculated based on significant level of 5%. 3.8 Classical Assumption Test Multicolinearity Test One of the classical assumptions in linear regression model is no multicolinearity in the independent variables. Multicolinearity is the degree of correlation between the independent variables, it is exist when two or more of the independent variables used in regression are correlated. Multicolinearity test is to prove whether the degree of correlation found in the regression model. A high degree of multicolinearity produces

7 26 unacceptable uncertainty (large variance) in regression coefficient estimates. Therefore multicolinearity is not allowed in the multiple regressions. To detect whether multicolinearity occurs in regression model, can be checked as follows: Variance Inflation Factor (VIF) and Tolerance (Gujarati, 1995) A regression model free of multicolinearity if: 1. Have VIF less than 5 2. Have tolerance value close to or equal Autocorrelation Test One of the basic assumptions of the regression model is the independence of the errors. This assumption is often violated when data are collected over sequential periods of time because a residual at any point in time may tend to be similar to residuals at adjacent points in time. Such a pattern in the residuals is called autocorrelation. When substantial autocorrelation is present in a set of data, the validity of a regression model can be in serious doubt (Levine, 2005 p533). Autocorrelation can be detected and measured by using the Durbin-Watson statistic. This statistic measures the correlation between each residual and the residual for time period immediately preceding the one of interest.

8 27 Table 3.3 Durbin-Watson Decision Making (Levine et al, 2008) Correlation If D-W statistic value No D > d U Yes D < d L Uncertain d L < D < d U Note: D d L d U = Durbin-Watson statistic D = Durbin-Watson lower = Durbin-Watson upper Heteroscedasticity Test The other assumption of the regression model is homoscedasticity. Homoscedasticity requires that the variation around the line of regression be constant for all values of the independent variable. This means that the errors vary the same amount when the independent variable is a low value as when it is a high value. The violation of this called heteroscedasticity. Heteroscedasticity can be detected from a plot of the residual with independent variable. To observe a case in which the heteroscedasticity occurs, examine the hypothetical plot of the residual with independent variable. If the plot appears to be a fanning effect in which the variability of the residuals increase as the independent variable increase, this is the indication that heteroscedasticity occurs. Otherwise, there is no apparent violation in the assumption of equal variance at each level of the independent variable.

9 Data Interpretation Statistical Package for the Social Science (SPSS 17) is used to interpret data. SPSS is among the widely used program for statistical analysis in social science. The base feature of this software is statistical analysis, data management and data documentation.