 Based on the provided data, several root causes for the performance issues can be identified:

1. Rejections at various stages by supervisors and employees: Many declarations were rejected multiple times by both supervisors and employees, causing delays in the approval process and affecting the overall performance. This suggests a need for improved communication, clarity of expectations, or potentially additional training for those involved in the approval process.

2. Administrative rejections: A significant number of declarations were rejected by administrators, leading to repeated submissions and lengthier processing times. This could indicate issues with the quality of the declarations, inconsistencies in administrative standards, or a high volume of declarations for the available resources.

3. Missing information: One declaration was rejected because it lacked necessary information, requiring resubmission. Ensuring that all necessary data is provided upfront could help reduce delays and improve performance.

4. Budget owner rejections: Several declarations were rejected by budget owners, which contributed to the lower performance for some request payments. This may indicate a need for better coordination between the budgeting and approval processes or more clarity about the criteria for budget approval.

5. Frequency of payments: The frequency of payments seems to have an impact on performance. Declarations associated with less frequent payments (e.g., once every 8 or 10 days) generally had lower performance figures compared to those with more frequent payments (e.g., daily). This could suggest that a higher volume of declarations can strain the approval process, leading to delays and reduced performance.

6. PRE_APPROVER rejections: Declarations were rejected at this stage multiple times, suggesting inconsistencies or unclear expectations for this level of approval. Improving the training or clarifying the role of the pre-approver might help reduce the number of rejected declarations and improve performance.

In summary, the root causes for the performance issues in this process include insufficient communication, quality control issues, missing information, inconsistent standards between roles, and a potentially high volume of declarations for the available resources. Addressing these specific concerns may help improve overall performance.