A challenging landscape calls for an evolution of the credit risk function
A bank’s bread and butter of generating high returns and increasing profitability is proving difficult in the current economic environment. At the same time as revenues are being challenged by low interest rates there is growing competition from new entrants and stringent capital and liquidity requirements. The business demands upon credit risk functions are increasing.
There is, however, an opportunity for credit risk functions to evolve from protecting the bank to generating growth. Credit risk usually takes up the bulk of Risk-Weighted Assets (RWA) for a bank1 (refer to Figure#1)and is therefore a sensible starting point to drive capital efficiencies. Improved RWA management can release capital that can be redeployed to fuel growth. For example, a £1bn saving in RWA redeployed at 2% Return on RWA (RoRWA) would enable an additional £20m profit per year.
Figure 1: RWAs by risk type (total EUR11.1tn as at Jun 2018)
There is more than one way to realise these benefits (process streamlining and advanced modelling are two that spring to mind), yet this article will focus on how these approaches are underpinned by enhancing risk data.
Dynamic and insightful analytics drive decision making
High data quality is fundamental to credit risk management; from risk assessments to modelling, risk MI, insights and decisioning. A lack of confidence in data quality leads to conservatism in calculations and holds back development and maintenance of advanced risk modelling, another key enabler for optimising RWA levels and delivering capital efficiency.
It’s not about big data – it’s about accurate and functioning data
Increasing amounts of data from a wider range of sources, alongside sophisticated analytical tools, can improve risk assessment/predictions, data reporting, pricing and product offering to consumers. But this is all dependant on completeness, validity, accuracy, consistency, availability and timeliness of data. To reap the benefits of new data sources, banks can explore how best to conduct a PoC with consumer permissioned data via Open Banking (e.g. realising benefits for mortgage applications); alternative data usage (e.g. utility/phone bills) and joint efforts for credit risk related data pooling and benchmarking activities.
We would recommend the following steps:
- Complete a data assessment against RWA and credit risk datasets to identify clear data requirements and data quality across competing demands in Credit Risk and Finance functions
- Define a data remediation strategy to increase and maintain accuracy and availability of data required for data-driven decisions and RWA savings
- Carry out a data sourcing assessment of current data source limitations and potential usage of new internal/external sources, including third party and non-traditional data sources
- Ensure strong data governance that aligns with Bank wide and CDO strategic objectives. Increased reliance on data and new regulations have increased the need to focus on data usage, privacy and security
Tools for automation and digitisation – benefits and risks
Banks can enhance their data capture and analytic capabilities in addition to reducing operating costs by embracing modern data technologies, such as Big Data, machine learning and Cloud Computing. Furthermore, banks are also exploring different use cases for emerging technologies, such as pilots/PoC for machine learning in credit risk. The most common use case thus far has been in credit scoring and decisioning, though with some application in capital and provisioning and stress testing3. Nevertheless, these new technologies are no panacea.
Emerging technologies have both the potential to solve a number of issues and create opportunities for banks; but first however, these issues need to be fully defined and clear requirements need to be considered before applying the following steps:
- Gain awareness of challenges to differentiate promise and peril against emerging tech. For example, machine learning’s ability to consume vast amounts of data to uncover patterns has sparked interest in the credit risk industry. Lenders should not only be concerned with predictive results of machine learning solutions, but also with the “, accuracy, bias and ethics behind the decisions and the intensified regulatory scrutiny.
- Complete a technology readiness assessment, including data and technology constraints (e.g. poor/unavailable datasets and legacy systems), innovative culture and clear understanding of feasible and measurable business cases. Approach should be defined across in house, open sourced and/or vendor applications.
- Conduct Proof of Concepts (PoC) for new technologies. Risk managers want to leverage these technologies for fully fledged solutions, but between data/systems constraints and significant investment of time and money, it can be difficult to justify. To combat this challenge, the definition and gradual implementation of PoC is crucial to deliver specific and measurable business value.
Meaningful metrics for actions, not just data visualisation
Banks should not underestimate the growing importance of advanced analytics in staying competitive. Analytical insight can support robust customer differentiation and risk decision capabilities, such as improved risk-based pricing, targeted segmentation, new product development and capital allocation.
We recommend the following steps:
- Outline measurable benefits to ensure efficient realisation of RWA savings and risk data enhancements. Key metrics should be defined and regularly monitored, as well as improving alignment of reporting across business, risk and finance
- Empowering insights by developing timely, dynamic, relevant and understandable MI and reports; underpinned by data quality, model standardisation and the latest visualisation tools
How far have credit risk functions got to go?
RWA reductions are not a new area of focus for banks. In fact, five UK large lenders reported lower Risk-Weighted Assets for end-2018 compared with a year ago2.
We have seen above how capital efficiencies can be achieved by increased data quality/availability, which in turn unlocks the potential for advanced risk modelling and better risk management. As a result, a better understanding of risks can facilitate business growth through more targeted risk based pricing and data driven strategic decisions, fed by timely and insightful analytics.
In spite of all of this however, the growing appetite for rich, high quality data is very far from being sated and most banks continue to struggle with significant challenges. On the other hand, that means a significant opportunity for those who move forward in a focused way.