Meeting EBA Guidelines with Advanced Analytics

As surely as the sun rises, central banking authorities will expand regulations on the retail lending sector. In this edition, we see the European Banking Authority laying out guidelines aimed at reinforcing banks’ portfolios and protecting consumers from loans they are unable to repay.

Published in May 2020, these guidelines will be implemented over a phased, three-year period, and aim to improve stability and financial health for both lenders and borrowers. Given the scope and depth of the new regulations – revamping credit reviews, pricing frameworks, and portfolio monitoring – it is imperative that banks begin their journey now.

Unsurprisingly, the EBA’s updates include additional recommendations around assessing creditworthiness – preventing defaults protects both bank and borrower. More notably, there is an intensifying focus on the role of pricing in protecting balance sheets. In fact, the EBA is now requiring lenders to develop pricing frameworks that consider not just loan-level risk but also tie pricing strategies directly to quantitative, risk-weighted performance metrics at the portfolio level. Analytically driven pricing can indeed be used to manage to metrics like Return on Risk-Weighted Assets (RORWA), but it requires a suite of interconnected models and calculations working in concert.

Connecting Risk and Profitability

The first hurdle in this process is connecting risk and profitability at the individual loan-level. While price is a key determinant of revenue, and default models give us a decent view of risk, connecting these pieces into a holistic view of profitability is non-trivial. Both revenue and risk must be calculated and discounted over the full life of the loan, integrating default and prepayment forecasts alongside potential shifts in the yield curve. Likewise, variable and fixed expenses for both origination and dealer compensation must be factored in as well. Most banks have the basic pieces for these individual calculations, but communicating inputs/outputs across systems and siloed quantitative teams is a difficult, time-consuming, and error-prone process.

The first step in streamlining these calculations is an integrated analytical system that incorporates all the necessary models and calculations in one place. Many banks have attempted to do this with ad hoc, home-built solutions. However, when your default models are built-in Python, prepayment is calculated using SAS, and NPV calculations are stored in a monster Excel sheet, an integrated, purpose-built system makes this integration far more manageable.

Designing and Testing Portfolio-Level Strategies

If we were only interested in the profitability of a single loan, we could stop there. But aligning pricing with portfolio-level considerations like RAROC further complicates the analytical picture. Varied loan terms, promotional offers, and market dynamics must all be incorporated. Likewise, the vast majority of lenders group loans together into pricing buckets, and thus are unable to tailor their pricing at the individual level – finding the right price to apply to hundreds or even thousands of diverse loans requires substantial analytical and computational horsepower.

Here, the integrated pricing analytics engine moves from ‘important’ to ‘mission-critical.’ The colocation and coordination between cost, risk, and revenue models is complex enough at the loan-level – harnessing these calculations for portfolio-level simulations is nearly impossible without a dedicated pricing analytics engine. Likewise, simulations involve large volumes of loans approved at hundreds (or thousands) of price points, demanding purpose-built, cloud-based computing power.

Wrap It All Up in Documentation

Of course, creating the pricing structure is not the end of the journey. Banks must then be able to demonstrate to the EBA and other regulators that what they have done is sound, safe, and fair. Lenders relying on a patchwork of systems for analytics and deployment will quickly find themselves inundated with mismatched reports and metrics spread across multiple teams.

Instead, a safer, simpler process uses a single system capable of both calculating the pricing structure as well as deploying that pricing structure to downstream sales channels. Records can be kept in one place, with systemization ensuring that there are no disconnects between steps.

Next-generation Pricing

While the transformations laid out in the EBA guidelines are oriented towards risk management, they also offer a tremendous opportunity to improve revenue and streamline operations as well. Employing analytical pricing engines for simulations also enables pricing optimization, which significantly increases both profitability and strategic agility. Likewise, end-to-end pricing tools can quantify and leverage relationships between products, improving personalized product bundles and customer retention across the Enterprise.

Fortunately, solutions like Earnix offer a ready-made, battle-tested way forward with data analytics in banking. Analytical components can be plugged into an AI-enabled pricing engine designed specifically for the task. Once optimized prices have been calculated, they can then be deployed directly to sales channels via OpenAPI, eliminating costly and time-consuming pricing execution processes. Origination data is then fed back into the system for a self-learning, automated system that constantly adjusts to changes in market, consumers, and competitors, with documentation stored at every step of the way.

The EBA has clearly presented lenders with a challenge. But those who rise to this challenge will realize not just reinforced risk profiles, but also significant improvements in revenue. Those with the right tools can begin seeing that money roll in before the first compliance reviews come due.