Regression Analysis of Positive Score in German Firms, Schemes and Mind Maps of Ethnology

Download Regression Analysis of Positive Score in German Firms and more Schemes and Mind Maps Ethnology in PDF only on Docsity! 176 The dependent variable is the Ethical Score (ES), which is measured as the difference of the Positive Score (PS) and the Negative Score (NS). The independent variables are defined as follows: the logarithmised revenue [in bn $] (logRevenue), the logarithmised market capitalization [in bn $] (logMarketCap), the logarithmised number of employees (log#ofemployees), the Net Profit Margin (NetProfitMargin), the Debt-to-Equity Ratio (Debt-to-Equity), the logarithmised Total Assets [in bn $] (logTotalAssets), and the industry as a dummy variable. Regression 1 (Reg1) uses the data from 2010 for both, the dependent and the independent variables. Regression 2 (Reg 2) uses the data for 2017 for both, the dependent and the independent variables. Regression 3 (Reg 3) uses the data from 2017 for the dependent variable and from 2016 for the independent variables in order to check for a time lag. Regression 4 (Reg 4) uses the average of the data from 2010 and 2017 in order to check the robustness. *** Significant at 1%, ** Significant at 5%, * Significant at 10%. Positive Score (PS) When using the Positive Score as the dependent variable, a very similar picture to the Ethical Score is depicted. The revenue also has positive coefficients for all four regressions and is significant on a 1% level for regression one and four, and significant on a 10% level for regression 2. This indicates that a high revenue has a positive impact on the Positive Score with data and time lag robustness. The market capitalization also shows a very similar regression result in comparison to the Ethical Score regression. All coefficients are under one and positive and only the second regression, that uses the 2017 data, is significant on a 10% level. This implies that the market capitalization has a positive impact on the Positive Score but is less significant than the revenue. The number of employees again has a negative but close to zero coefficient for the first regression that uses the 2010 data and for the other three regressions has positive coefficients. It is significant on a 1% level when using the 2017 data and the robustness check for time lag is significant on a 5% level. The number of employees appears to have a positive impact on the Positive Score in more recent years and appears to not have that effect earlier. The Net Profit Margin has all negative coefficients that are close to zero. They are significant on a 1% level for regression two to four and significant on a 10% level for regression one. The Net Profit Margin appears to have no to slightly negative impact on the Positive Score. The same can be found for the Ethical Score. As the results for regression three and four are similar to one and two, data and time lag robustness can be assumed. The Debt-to-Equity Ratio regression results for PS are also similar to the ES. The coefficients are all close to zero and a have a negative sign. The only difference is that the coefficient for the first regression is significant on a 10% level implying that a high Debt-to-Equity Ratio has a small, negative impact on the PS. The Total Assets coefficients are also all close to zero with a negative sign. None of them are statistically significant while the data is robust. 177 The F-Score is positive and statistically significant on a 1% level indicating a good regression model fit. As the results for regression three and four are similar to the results from one and two, a robustness of the data and time delay can be assumed. Table 40 - Regression Results Germany PS GER_PS Reg1 t10 → t10 Reg2 t17 → t17 Reg3 t17 → t16 Reg4 tØ10,17 → tØ10,17 Coefficient p-value Coefficient p-value Coefficient p-value Coefficient p-value Intercept 9.5406*** 0.0000 9.7675*** 0.0000 10.3410*** 0.0000 9.7019*** 0.0000 logRevenue 2.0396*** 0.0008 0.6158 0.3857 1.2190* 0.0872 1.6782*** 0.0092 logMarketCap 0.6736 0.2134 0.8561* 0.0976 0.5188 0.3357 0.7391 0.1525 log#ofemployees -0.0270 0.9461 1.3641*** 0.0036 1.1267** 0.0155 0.5546 0.1856 NetProfitMargin -0.6364* 0.0568 -0.2511*** 0.0055 -0.2598*** 0.0015 -0.3739*** 0.0033 Debt-to-Equity -0.5191* 0.0964 -0.0644 0.7853 0.0019 0.9900 -0.3038 0.3295 logTotalAssets -0.4136 0.3510 -0.2754 0.5634 -0.4876 0.2305 -0.5279 0.2740 Industry -0.0684 0.5643 0.0421 0.7252 0.0076 0.9492 0.0021 0.9847 R-Squared 0.4107 0.3830 0.3884 0.4247 F-Score 11.2500*** 10.0217*** 10.2507*** 11.9179*** Significance F 0.0000 0.0000 0.0000 0.0000 Table XY – GER_PS Regression. The table presents the regression results of the Positive Score (PS) and the independent variables indicating the size of the firm in Germany. The dependent variable is the Positive Score (PS), which is the sum of all positive scores in the screening. The independent variables are defined as follows: the logarithmised revenue [in bn $] (logRevenue), the logarithmised market capitalization [in bn $] (logMarketCap), the logarithmised number of employees (log#ofemployees), the Net Profit Margin (NetProfitMargin), the Debt-to-Equity Ratio (Debt-to-Equity), the logarithmised Total Assets [in bn $] (logTotalAssets), and the industry as a dummy variable. Regression 1 (Reg 1) uses the data from 2010 for both, the dependent and the independent variables. Regression 2 (Reg 2) uses the data for 2017 for both, the dependent and the independent variables. Regression 3 (Reg 3) uses the data from 2017 for the dependent variable and from 2016 for the independent variables in order to check for a time lag. Regression 4 (Reg 4) uses the average of the data from 2010 and 2017 in order to check the robustness. *** Significant at 1%, ** Significant at 5%, * Significant at 10%. Negative Score (NS) When using the Negative Score as the dependent variable, only two independent variables have a statistically significant impact, namely the Revenue and the Total Assets. The revenue has all positive coefficients but close to zero. As the coefficients are all similar, time lag and data robustness are assumed. The coefficient for the first regression is statistically significant on a 5% level, while the coefficients for the third and fourth regression are statistically significant on a 10% level. The results imply that the revenue has small, positive impact on the Negative Score. This is counterintuitive as the revenue also has a positive impact on the