The Effect of Liquidity and Profitability on the Capital Structure as the Moderator Variable of Subsector Retail Companies Listed on Indonesian Stock Exchange

A company that operates in parallel with its development always needs additional capital. When a company is built, the owner can decide which capital source is going to be used whether it is only from common index capital or long time liability. Every decision that is made on the capital source always has its effect. Capital structure is the comparison between liabilities and the company’s capital. One of the important matter that often faced by the manager of a company is to decide on the right balance between liabilities and capital. The following is the condition of the capital structure which is being projected with the debt to equity ratio of the subsector retail companies listed on the Indonesian Stock Exchange in the period of 2011-2015. Abstract

It is seen from the table above that debt to equity ratio of the company with JKSW stock symbol had a significant decrease compared to the other companies' stock symbols. This shows that the profit that the company gained was having a decrease which caused the investor in the stock market did not add any fund to their issuer. Therefore, this will affect the company such as lowering the company's performance in achieving its goals.
The purposes of this research are: Firstly, it is to test and to analyze whether the current ratio will affect the Return on Assets with debt to equity ratio as the moderator variable of the subsector retail companies listed on Indonesia Stock Exchange in the period of 2011 -2015. Secondly, it is to test and to analyze the cash turnover that will affect the Return on Assets with debt to equity ratio as the moderator variable of the subsector retail companies listed on Indonesia Stock Exchange in the period of 2011-2015. Thirdly, it is to test and to analyze the size of the company which will affect Return on Assets with debt to equity ratio as the moderator variable of the subsector retail companies listed on Indonesia Stock Exchange in the period of 2011 -2015. Fourthly, it is to test and to analyze the current ratio, the cash turnover and the company size that will affect the Return on Assets with debt to equity ratio as the moderator variable of the subsector retail companies listed on Indonesian Stock Exchange in the period of 2011-2015.

Problem Formulation
The problem formulations of this

Cash Turn Over
According to Kasmir (2012:140), cash turnover has the function to measure the level of working capital required by the company to pay bills and the finance sales. Which means, this ratio is used to measure the level of cash availibilty to pay the bills (debt) and the other costs that are associated with the sale.

Company Size
According to Najmudin (2011: 316), company size is that generally, the large scale company is easier to obtain debts compared to the small companies because of the trust level given by the creditor to the big companies. According to Halim (2015: 125), the larger the size of a company, the greater also the tendency of a company to use foreign capital. This is because the large company needs large funds too to support their operational and one of the alternative fulfillment is to get foreign capital if the company owner's capital is insufficient.

Return on Assets
According to Husnan (2006: 73), Return on Assets can also be said as the profitability of the company which is a ratio that measures how much profit owned by the owner of the capital. According to Murhadi (2013:64), Return on Assets is the reflection of how much the return generated for the shareholders over each dollar of money is being invested. According to Ichsan (2021) Return on Asset (ROA) is one form of profitability ratio, by using after various capital costs and total assets owned by banks, it can see the ability of a company to be able to earn profit / profit. Because, return on assets is a measurement tool used in the ability of the company and assess the effectiveness to get profit and profit.

Debt to Equity Ratio
According to Jusuf (2088: 55), debt to ratio is the ratio between total liabilities to the total of equity. This ratio shows the extent of the capital to guarantee all of the liabilities. According to Harahap (2013:303), debt to equity ratio illustrates the extent to which the owner's capital can cover the debts to outsiders.

Previous Research
Yuke (2005), did a research with the title "Factors that Affect the Capital Structure of Go Public Manufactures in Jakarta Stock Exchange". The result of research showed that partially, the company size had a significant effect towards the capital structure and the profitability had the significant positive effect towards the capital structure. While simultaneously, the company size, business risk, asset growth, profitability and corporate ownership structure affect the Capital Structure.
Thomi Irvan (2016), did a research with the title "The Effect of Profitability and Liquidy on the Capital Structure of Insucrance Companies Listed on Indonesia Stock Exchange (IDX) in the Period of 2012 -2014". The partial research result showed that the variables of profitability and liquidity did not affect DER. While simultaneously, profitability and liquidity did not affect DER.

III. Research Methods
This research tends to be correlational which explain the association between the research's variables. Using secondary data such as documents and financial reports that are related to this research. The data is obtained from Indonesia Stock Exchange website, derived from the annual financial report to be sampled in the period of 2011-2015 which is downloaded from the official website of Indonesia Stock Exchange namely www.idx.co.id.
The population in this research are the 16 subsector retail companies listed on Indonesia Stock Exchange in the period of 2011-2015. Using purposive sampling, 10 out of 16 companies are selected to be the samples. The technique of the data collection in this research uses documentation study, that is by downloading the financial report of the subsector retail companies listed on Indonesia Stock Exchange in the period of 2011-2015. Moderator Variable is a variable that strengthens or weakens the correlation between one variable to another.
There are 2 equation models in general, which are: 1. First Method Moderation regression equation with iteration test: Y = a + b1 X1 + b2 X2 + b3 X1X2 + e Description: Y = Profitability A = Constant b1 = Regression coefficient for cash turnover b2 = Regression coefficient for company size b3 = Regression coefficient for moderator X1 = Cash Turnover X2 = Company size X1.2 = Interaction 2. Second Method a. Absolute Difference Value. This is done by finding the difference of absolute value standardized between two independent variables. b. If the absolute difference value between two independent variables are positive significant, that means the variable moderates the correlation between the independent variable and the dependent variable.
Below is the research model framework:

Figure 1. Conceptual Framework
The analyses of this research data are using the first two analyses, by using the descriptive analysis and statistical analysis such as multiple linear regression analysis.

Statistical Analysis
Before doing the statistical analysis by using multiple linear regression analysis, it is important to fulfill the requirements of classical assumption test. The following is the explanation of the classical assumption test:

a. Normality Test
Normality test is done in order to test whether the disturbing variable or the residual variable has a normal distribution in the regression model (Ghozali, 2011). There are two methods to detect whether the data is normally distributed or not by using graph analysis (scatterplot), and statistical analysis (Kolmogrov-Smirnov).
The following is the result of classical assumption test based on the normality test by using Kolmogrov-Smirnov: Based on the table of One sample Kolmogrov-Smirnov, it is seen that the value of Asympg.Sig (2-tailed) is above 0,05 or above 5%. This shows that the data is normally distributed.

b. Heteroscedasticity Test
Heteroscedasticity Test purpose is to test whether there is a variance inequality of the residual in one observation to the another observation in the regression model (Ghozali, 2016: 134-138). In order to see or to detect the occurrence of heteroscedasticity in this research, the researcher uses the scatterplot chart.

Figure 2. Heteroscedasticity Test (First Regression Model)
Source: Result of Data Processing, 2020

Figure 3. Heteroscedasticity Test (Second Regression Model)
Source: Result of Data Processing, 2020 From the scatterplot chart, both outputs appear to have dots spreading randomly either from above or below the number zero (0) on the Y axis. The dots do not gather in a place. Thus from the scatterplot chart, it is concluded that there is no heteroscedasticity happened in the regression model.

c. Multicollinearity Test
The purpose of Multicollinearity Test is to test whether there is a correlation between independent variables in the regression model (Ghozali, 2016: 103-104). Detecting the occurrence of multicollinearity in the regression model can be done by looking at the tolerance and the variance inflation factor with the following assumptions: Based on the table of multicollinearity test, the output of the first regression shows that the tolerance value is below 0.05 and the VIF value is above 10. Thus, it can be indicated that the symptom of multicollinearity occurs which causes the researcher performs the second test as a moderator or amplifier variable for the following multicollinearity test: It is seen that on the table above, the output in second regression involves the moderator variable. Therefore, there is no symptom of multicollinearity occurs in this research.

d. Autocorrelation Test
According to Ghozali (2016: 107 -108), the purpose of autocorrelation is to test whether there is a correlation between the error in the t period and the error in t-1 period (before). If the correlation occurs, then it is named problem autocorelation. The method used to detect the occurance or the absence of autocorrelation is by using Durbin Watson test (DW test).

Coefficient of Determination (R2)
The purpose of the coefficient of determination is to see how far the ability of a model in explaining the dependent variable According to the table of Model Summary, the value of R square is 0.243 and after using the moderator variable, the value of R-square becomes 0.226. Thus, the residual of both R-square values outside this research and the strengthening of the number are seen after the moderator value is inserted in this research.

Group Test (F Test)
The F Test purpose is to test if the variable of cash turnover and company size affect the Return on Assets in the first regression model or whether the variables of the cash turnover and the company size affect the Return on Assets simultaneously if there is capital structure moderator variable in the second regression model. Both of the model results can be seen below: Based on the output of the second regression above, it is seen that the significant value is less than 0.05. Thus, the capital structure moderates both of the variables in this research.

Individual Test (t Test)
The purpose of t Test is to test whether the cash turnover variable and the company size affect individually or partially on Return on Assets in the first regression output or whether the variables of the cash turnover and company size affect partially on Return on if there is capital structure moderate variable in the second regression output. Both of regression outputs can be seen as below: Based on the research result, it is shown that partially the value of t arithmetic is -1.016 and t table is 2.228, which means t arithmetic < t table. Meanwhile, the significant value is 0.315 which is bigger than 0.05. This means the capital structure does not moderate both variables in this research.

b. Second Regression Model
Based on the partial test result, t arithmetic is 1.626 with a significant value of 0.111, the value of t table is 2.228. This means t arithmetic < t table with the significant value of 0.111, which is bigger than 0.05. Therefore, it is concluded that capital structure does not moderate both of the variables in this research.
On the other hand, this research is not in line with Yuke (2005) and Nadzirah (2016). The partial research result showed that company size and profitability has a positive and significant effect on the capital structure.

IV. Conclusion
The conclusion of this research is that the first and the second regression models simultaneously show the variable of capital structure affects and moderates both of the variables in this research. While partially, both of the regression models in this research show that capital structure does not moderate both of the variables in this research.