CRR model optimisation
Breaking down clusters of apparently homogeneous customers and building robust risk profiles to focus monitoring on relevant-risk customers.
Need to align CRR model with requirements
Regulatory requirements and risk coverage
Our clients need to enhance their AML/CTF monitoring effectiveness, taking into consideration the existing obligations deriving from the AML/CTF regulatory framework, international best practices, as well as those deriving from the EBA Guidelines and the new “AML Package” as their CRR models have been in operation for many years without any thorough review and complex gap assessment against the current expectations.
Resources spent on EDD
Frequently, there is a significant percentage of customers assigned to the medium- or high-risk level which is often caused by a couple of static risk indicators having an exaggerated impact on inadequately high number of customers. A lot of time and resources are spent on reviewing customer files with rather usual risk profiles.
Targeting relevant customers
CRR models often overdepend on a few regulatory risk indicators with a strictly bipolar scoring. A customer is directly assigned with either medium/high risk, or no risk based on their political exposure, geographical links, business activity, etc. But there is nothing much in between. The other aspects, especially behavioural indicators, are usually under represented. See the typically observed aspects below.
Frequent suboptimal aspects, we can help you address
Unexpectedly high percentage of non-retail customer in medium-risk level: It is necessary to analyse the major contributors, apply more granular scoring.
Only a couple of risk indicators contribute to risk profile of retail customers: The parameters of the understated contributor need to be tuned to reflect statistical distribution of customer data.
Behavioural risk indicators are missing or having only a negligible contribution: (New) indicators can be configured to score customers based on the size of deviation from usual behaviour
Gaps between requirements and CRR model: It is necessary to validate if all external and internal requirements are accurately translated into the CRR model business definition.
Scoring lists do not sufficiently diversify the risk associated with individual categories: Applying more granular weights and/or more detailed categories is essential for efficient scoring.
Deviations between model business specification and actual implementation: Program code and parametrization review, or back-testing validation, can reveal any typos or incorrect parameters.
Static risk indicators with an exaggerated impact: The indicator‘s rules need to be customized to correspond with the statistical distribution of customers‘ data, providing more detailed categorization.
Are you unsure how your CRR model work?
Check yourself with our self-service tool!
Contacts
Senior Manager, Financial Crime Advisory, PwC Czech Republic
Tel: +420 703 185 657
You are looking for an expert to help you; you want to request our services; or simply ask something? Let us know about yourself and we will get back to you as soon as possible.