Retail is an important link in the distribution process of goods and is the last link in a distribution process. Through retail, a product can meet directly with its users. The retail industry here is defined as an industry that sells products and services that have been added value to meet the needs of individuals, families, groups or end users. Communities in big cities and small towns have now been facilitated by the presence of shopping centers. The location is also easy to find. This condition is very beneficial because people just have to choose which outlets to enter. These outlets include various types of retail businesses. The development of the retail business in Indonesia is quite fast lately, especially modern retail in all its variations.

Several factors that support the development of modern retail businesses include large market opportunities, the development of manufacturing businesses that will supply their products to retailers (retailers) and the government's efforts to encourage economic growth by developing a retail business. As with other companies, the main goal of retail companies is to generate the maximum profit from various assets used as company capital to drive various company policies. The company's ability to generate profits is known as profitability. Profitability has an important meaning for the company because it is one of the bases for assessing the condition of a company. The profitability of a company shows a comparison between profits with assets or capital that generates these profits. In other words, profitability is the company's ability to generate profits over a certain period. Profitability can be projected with Return on Assets (ROA) because it can show how the company's performance is seen from the use of all assets owned by the company in generating profits.

The performance of retail companies listed on the Indonesia Stock Exchange differs from one company to another, where there are frequent fluctuations in the company's performance in generating profits. Sawir explains that increasing company profitability is very dependent on the efficiency or effectiveness of the company    in    utilizing    all   resources   or   assets   (assets) owned by a company. One way to measure the efficiency and effectiveness of the company in using the company's assets is the ratio of total assets turnover.

Leverage conditions as measured by a high Debt to Equity Ratio can increase profitability due to greater availability of funding. Debt to equity ratio is a ratio that shows the ability of the company's own capital to meet all of its obligations. Hanafi argues that a high debt to assets ratio indicates the company uses high debt. The use of high debt will increase funding sources which if used efficiently will increase profitability, on the other hand, high debt will also increase risk. This shows that working capital and debt to equity ratio have an important role in influencing the profitability of the company.

Sufficient working capital will provide benefits for the company. With sufficient working capital, the company in financing all company operations in order to increase the company's profit (profit) to the maximum. The company's working capital can be measured by working capital turnover, where Riyanto explains that working capital turnover is a ratio used to test the effectiveness of the use of working capital, namely the ratio between sales and working capital. The higher the net profit margin also indicates the higher the net profit (profit) obtained by the company so that the net profit margin ratio will affect all other profitability ratios, including the ratio of return on assets, return on equity, gross profit margin and operating profit margin.

The research approach is quantitative research. Based on the level of explanation, this research is an associative research. Associative research or causal research (causal relationship) is research that wants to see whether a variable that acts as an independent variable has an effect on other variables that are the dependent variable. The population in this study are all retail companies listed on the Indonesia Stock Exchange until 2020, totaling 25 retail companies. This study uses a purposive sampling technique which determines a sample from the population with certain criteria. The population that can be used as a sample is a population that meets the following criteria:

• Retail companies listed on the Indonesia Stock Exchange (IDX) during 2014-2019

• Retail companies publish company data consisting of Revenue, Total Assets, Debt to Equity Ratio (DER), Current Assets, Current Liabilities, Net Profit Margin (NPM) and Return on Assets (ROA) in 2014 – 2019

• Data published by retail companies from 2014-2019 has been audited

Based on these criteria, the number of population that meets the requirements to be used as a sample is 17 companies, of which 8 other companies do not meet the requirements. So that the companies that can be taken as samples are 17 retail companies listed on the Indonesia Stock Exchange. The data collection technique used in this research is a document study, by studying data from documents obtained from a summary of the company's recorded financial performance and the company's annual report from 2014 to 2019. This study uses panel data regression analysis techniques.

In the panel regression method, there are three choices of models that can be used, namely Common Effect Modal (CEM), Fixed Effect Model (FEM) or Random Effect Mode (REM). These three models have different criteria and researchers must be able to choose the most appropriate model based on the research data that the researcher has. Therefore, three tests are needed to determine which model is the most appropriate for the researcher to use. The three tests are Chow Test, Hausman Test and Lagrage Multiplier Test. If two of these three tests decide on the same model, then that model is the most suitable for the researcher.

The results of model testing using Chow test data are shown in Table 1.

Table 1: Chow Test Results

The statistical distribution value of the chi-square in the Table 1 based on the results of calculations using Eviews 9 shows that the probability F in the Chow test has the effect of Total Assets Turnover (X1), Debt to Equity Ratio (X2), Working Capital Turnover (X3) and Net Profit Margin (X4) to Return on Assets (Y) is 0.0000. This value is smaller than 0.05 or 5%, statistically H0 is rejected and H1 is accepted, so the correct model used is the Fixed Effect Model (FEM). After that, it is necessary to do another test, namely the Hausman test to find out whether it is better to use the Fixed Effect Model (FEM) or the Random Effect Model (REM).

Hausman test is used to determine whether the most suitable modal is Fixed Effect Model (FEM) or Random Effect Model (REM). If the conclusion is accepted by Ho, then choose the Random Effect Model (REM), if the conclusion is accepted H1 then select the Fixed Effect Model (FEM). The results of model testing using the Hausman test can be seen in Table 2.

Table 2: Hausman Test Results

The Table 2 shows that the probability in the Hausman test on the effect of Total Assets Turnover (X1), Debt to Equity Ratio (X2), Working Capital Turnover (X3) and Net Profit Margin (X4) on Return on Assets (Y) is equal to 0.0000 The test results show a probability value of <0.05, so H0 is rejected and H1 is accepted. Therefore, the correct model used is the Fixed Effect Model (FEM) because H1 is accepted.

The results of the Chow Test and the Hausman Test both show that the most appropriate model is the Fixed Effect Model (FEM) so that there is no need for further testing with the Lagrange Multiplier Test because two of the three tests, namely the Chow Test and Hausman Test, have shown that the most The right one is the Fixed Effect Model (FEM). So in this study the model used is the Fixed Effect Model (FEM).

Hypothesis testing consists of the F test (simultaneous test) to find out how significant the influence of the independent variable on the fixed variable is simultaneously and the t test (partial test) to find out how significant the influence of the independent variable on the fixed variable is partially (Table 3).

Table 3: Results of t-Test (Partial Test)

The  results  of  the  t-test  indicate  that  the  t-count for   the   Total  Assets  Turnover  (X1)  variable  is 2.555630, with a t-table value of 1.985, it is known that the value, which is 2.555630 >1.985, rejects Ho and accepts Ha. The significant value t of the Total Assets Turnover (X1) variable is 0.0000, where this value is much  smaller  than  the  significant  threshold  of  0.05. Then  reject  Ho  and  accept  Ha.  Therefore,  the  test results meet the equation tcount> ttable and significant <0.05. So it can be concluded that partially there is a significant effect of Total Assets Turnover (X1) on Return on Assets (Y).

The results of the t test show that tthe count owned for the Working Capital Turnover (X3) variable is 0.004550, with valuettable of 1.985, it is known that the value of t-count<t-table, which is 0.004550 <1.985, then reject Ha and accept Ho. The significant value t of the Working Capital Turnover (X3) variable is 0.9760, where this value is much greater than the significant threshold of 0.05. Then reject Ha and accept Ho. Therefore, the test results do not meet the equation t-count>t-table and significant <0.05. So it can be concluded that partially there is no significant effect of Working Capital Turnover (X3) on Return on Assets (Y)

The results of the t-test indicate that the t-count for the Net Profit Margin (X4) variable is 35.30099, with a t-table value of 1.985, it is known that the t-count value > t-table, then reject Ho and accept Ha. The significant value t of the Net Profit Margin (X4) variable is 0.0000, where this value is much smaller than the significant threshold of 0.05. Then reject Ho and accept Ha. Therefore, the test results meet the equation t-count>t-table and significant <0.05. So it can be concluded that partially there is a significant effect of Net Profit Margin (X4) on Return on Assets (Y).

The results of the t-test indicate that Total Assets Turnover (X1) and Net Profit Margin (X4) are the variables that affect Return on Assets (Y) but Net Profit Margin (X4) is the variable that has the most influence on Return on Assets (Y) because it has the largest tcount value is 35.30099.

The F test (simultaneous test) was conducted to see the effect of the independent variables on the dependent variable simultaneously or together. The method used is to look at the level of significant where the reference point is a significant value of 0.05 or 5% or 5% or by comparing F-count to F-table.

The results of the F-test processed with the Eviews 9 application can be seen in Table 4.

Table 4: F Test Results (Simultaneous Test)

The results of the F-test from the Table 4 show that the F-count obtained is 106.3372. This F-count value is much greater than the F-table value which is 2.465. Then it can be concluded reject Ho and accept Ha. Based on the significant value produced, it is known that the significant value of the F test conducted is 0.0000 where this significant value is much smaller than 0.05. So reject Ho and accept Ha so that based on the F test it can be concluded that the regression model in this study is Total Assets Turnover (X1), Debt to Equity Ratio (X2), Working Capital Turnover (X3) and Net Profit Margin (X4) simultaneously. significant effect on Return on Assets (Y).

Based on the results of the tests and analyzes that have been carried out, some conclusions can be drawn as follows:

• Total assets turnover partially positive and significant effect on return on assets in retail companies listed on the Indonesia Stock Exchange

• Debt to equity ratio partially has no significant effect on return on assets in retail companies listed on the Indonesia Stock Exchange

• Working capital turnover partially has no significant effect on return on assets in retail companies listed on the Indonesia Stock Exchange

• Net profit margin partially positive and significant effect on return on assets in retail companies listed on the Indonesia Stock Exchange

• Total assets turnover, debt to equity ratio, working capital turnover and net profit margin simultaneously have a positive and significant effect on return on assets in retail companies listed on the Indonesia Stock Exchange

1. Alpi, M.F., and A. Gunawan. “The Effect of Current Ratio and Total Assets Turnover on Return on Assets in Plastic and Packaging Companies.” Journal of Axiom Accounting Research, vol. 17, 2018, pp. 1–36.

2. Atmaja, L.S. Financial Management Theory and Practice. Yogyakarta: ANDI Publisher, 2013.

3. Month, T.P.L. “Effect of Working Capital on Profitability Level at PT Adira Dinamika Multi Finance Tbk.” Journal of Management and Finance, vol. 4, no. 1, May 2015.

4. Brealey, M. Fundamentals of Financial Management. Jakarta: Erlangga, 2016.

5. Brealey, R.A., S.C. Myers, and A.J. Marcus. Fundamentals of Corporate Financial Management. Vol. 1, 5th ed., Jakarta: Erlangga, 2015.

6. Efendi, A.F.W., and S.S.A. Wibowo. “The Influence of Debt to Equity Ratio (DER) and Debt to Asset Ratio (DAR) on the Performance of Companies in the Financial Sector Listed on the Indonesia Stock Exchange.” Journal of Applied Managerial Accounting, vol. 1, no. 2, 2017, pp. 157–166.

7. Felany, I.A., and S. Worokinasih. “The Effect of Working Capital Turnover, Leverage and Liquidity on Profitability (Study on Food and Beverage Sub-Sector Companies Listed on the Indonesia Stock Exchange in 2012–2016).” Journal of Business Administration (JAB), vol. 58, no. 2, May 2018.

8. Ghozali, I. Application of Multivariate Analysis with IBM SPSS 19 Program. 5th ed., Semarang: Diponegoro University, 2015.

9. Halim, A. Business Financial Management: Concepts and Applications. Jakarta: Media Discourse Partners, 2015.

10. Horne, V., and Wachowicz. Financial Management Principles. Jakarta: Salemba Four, 2015.

11. Irawan, and Z.A.T. Silangit. Financial Statement Analysis. Medan: Smartprint Publisher, 2018.

12. Irawan, and M.D. Tuah. Research Methodology Applications and Data Processing Techniques. Medan: Smartprint Publisher, 2017.