Erasmus University Rotterdam Erasmus School of Economics Master Accounting, Auditing and Control Master's Thesis Accounting, Auditing & Control Successful-Efforts
Earnings model for the successful-efforts method
Download 0.83 Mb.
|
- Navigate this page:
- 6.4.1 Meeting assumptions
- 6.4.2 Regression model
Earnings model for the successful-efforts methodThe fourth model is also an earnings model. The variables used in this model are based on the disclosure of research and development expenditures using the successful-efforts method. The first section asses whether the assumptions are met. The second section provides the regression results. First the absolute model will be discussed and afterwards the scaled model. 6.4.1 Meeting assumptionsThe histograms show a normal distribution. There is some skewness but the sample exists of 38 observations. Some skewness is allowed by this sample according to the central limit theorem. The assumption of normality is met. There are some outliers in the scatterplots. Those outliers can be explained by the difference in size of the firms. The variables can be used in regression and the assumptions homoscedasticity en linearity are met. The scatterplots and histograms are placed in appendix three. The fourth assumption is non multicollinearity. In the next section in the coefficient tables this assumption is tested by the variance inflation factors. The VIF stay below 10, which indicate that the assumption of non multicollinearity is met. 6.4.2 Regression modelAfter the assessment of the assumptions the regression models are tested. First, the regression model with the scaled variables is tested. Second, the regression model with the absolute variables is tested. The regression model with the scaled variables provided an explanatory power of 0,188. The variables tested in this model are not significant. Perhaps the control variables give new inside. The SPSS output of the scaled model are showed below.
The control variables that are used testing this model are R&D intensity and code law/common law. These control variables don’t give new insight on the model. The variables used in this analysis stay insignificant. The explanatory power has increased minimum.
The second model for the earnings returns relation for the successful-efforts method is the model with the absolute values. The explanatory power is 0,883. This is a large increase in opposite to the scaled model. The variables used are also significant. Only the change in net income before R&D isn’t significant. The change in direct research and development expensed has a p-value of 0,06. Due to the sample size of 38 this variable is seen as significant. The variable net income before R&D has a positive effect on the returns. The change variable also shows a positive effect, meaning that an increase in net income before R&D has an extra positive effect on the returns. The variable R&D expenditures that are directly incurred and the R&D amortization also have a positive affect on the returns. Although, the mutation of those two variables have an extra negative effect on the returns. The SPSS outputs are provided beneath.
The control variable revenue and common law/code law is added to the model. The variable common law/code law is not significant in this regression model. Revenue is significant but it didn’t lead to a change in the relations between the other variables and the returns. The explanatory power increased from 88% to 90%.
The results from the earnings for the successful-efforts method and the cash-expense method will be compared and analysed in chapter seven “Analysis”.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||