Pool data flow
About this app
A central pool node connects to pilot banks, central method actor, central IT actor and supervisor. Particle flows, an edge click shows format, frequency, owner and audit question.
The idea
When several banks jointly develop and run an IRB rating system in a data pool, many data deliveries arise: default events and collateral data flow from the banks into the pool, methodology and software come from central units, and estimated parameters and reporting data go back to the banks and to the supervisor. For internal audit, each of these routes is a separate audit area: in which format, how often and under whose responsibility is data delivered, and how is data quality ensured?
What the app shows
A 3D scene with the pool in the centre, seven pilot banks in a ring, the supervisor above, and one central unit each for methodology (Z-Meth) and IT (Z-IT). Particles travel along the edges; inbound deliveries (bank → pool) and outbound ones (pool → bank or supervisor) have different colours. Five delivery classes can be shown or hidden: default data, collateral, methodology, software and RWA reporting. Clicking an edge opens details with format, frequency, owner and a typical audit question.
Explore it
Hide all classes except one and trace which actors are involved. The detail view lists an audit question for each delivery, for example on the completeness of default events or on the reconciliation between pool output and the general ledger. A typical opening question from audit is: which data leave the bank, and which come back? All details are generic examples and do not describe a real pool.
Frequently asked questions
- What is an IRB data pool?
- A data pool combines the credit risk data of several institutions so that a shared rating system can be developed and validated on a larger data base. This matters especially for smaller banks whose own default history would be too short for robust estimates. Each institution remains responsible for applying the system to its own portfolio.
- What does internal audit check in data deliveries?
- Typical audit areas are the completeness and accuracy of delivered data, format and frequency, clear responsibilities, a four-eyes check before sending, and reconciliation with source systems such as the general ledger. For methodology and software, versioning, approvals and change logs are added.
- Why does data lineage matter in regulatory reporting?
- Data lineage describes the path of a data item from its source to the report. Only if this path is known and documented can errors be traced back and controls be placed effectively. For audit, a path you do not know is a path you cannot check.
- What does the pool data flow app show?
- It shows a fictional IRB pool as a 3D network of pilot banks, central units and the supervisor. You can filter delivery classes and click any edge to see format, frequency, the responsible unit and a matching audit question.
Subject: IRBA audit (banking pool)
More from IRBA audit (banking pool)
A VisuApp by heyprof: interactive, free in your browser, no sign-up. What is a VisuApp? · All apps