With recent volatility in financial markets, firms need increasingly faster pre-trade and risk calculations to react swiftly to changing markets. Traditional computing methods for these calculations, however, are becoming prohibitively expensive and slow to meet the growing demand. GPU computing has recently garnered significant interest, with advances in the fields of advanced machine learning techniques and generative AI technologies, such as ChatGPT. Financial institutions are now looking at gaining an edge by using GPU computing to accelerate their high-dimensional and time-critical computing challenges. 

The MVA Computing Challenge 

The timely computation of MVA is essential for pre-trade and post-trade modelling of bilateral and cleared trading. Providing an accurate measure of future margin requirements over the lifetime of a trade requires the frequent revaluation of derivatives with a large volume of intensive nested Monte Carlo simulations. These simulations need to span a high-dimensional space of trades, time steps, risk factors and nested scenarios, making the calculation of MVA complex and computationally demanding. This is further complicated by the need for an increasing frequency of intra-day risk calculations, due to recent market volatility, which is pushing the limits of what can be achieved with CPU-based computing.  

An Introduction to GPU Computing 

GPU computing utilizes graphics processing units, which are specifically designed to handle large volumes of parallel calculations. This capability makes them ideal for solving programming challenges that benefit from high levels of parallelization and data throughput. Consequently, GPUs can offer substantial benefits over traditional CPU-based computing, thanks to their architectural differences, as outlined in the table below. 

A comparison of the typical capabilities of enterprise-level hardware for CPUs and GPUs.

It is because of these architectural differences that CPUs and GPUs excel in different areas: 

• CPUs  feature fewer but more powerful cores, optimized for general-purpose computing with complex, branching instructions. They excel in performing serial calculations with high single-core performance. 

• GPUs consist of a large number of less powerful cores and with higher memory bandwidth. This makes them ideal for handling large volumes of parallel calculations with high throughput. 

Solving the MVA Computational Challenge with GPU Computing 

The requirement to calculate large volumes of granular simulations makes GPU computing especially well-suited to solving the MVA computational challenge. The use of GPU computing can lead to significant improvements in performance for not only MVA but a range of problems in finance, where it is not uncommon to see improvements in calculation speed of 10 – 100x. This performance increase can be harnessed in several ways: 

• Speed: The high throughput of GPUs provides results more quickly, providing faster risk calculations and insights for decision-making, which is particularly important for pre-trade calculations. 

• Throughput: GPUs can more quickly and efficiently process large calculation volumes, providing institutions with more peak computing bandwidth, reducing workloads on CPU-grids that can be used for other tasks. 

• Accuracy: With greater parallel processing capabilities, the accuracy of models can be improved by using more sophisticated algorithms, greater granularity and a larger number of simulations. As illustrated below, the difference in the number of Monte Carlo simulations that can be achieved by GPUs in the same time as CPUs can be significant. 

The difference in the number of Monte Carlo paths than can be simulated in the same time between an equivalent enterprise-level CPU and GPU.

Case Study: Our approach to accelerating MVA with GPUs 

To illustrate the impact of GPU computing in a real situation, we present a case study of our work accelerating MVA calculations for a major bank. 

Challenge: A large investment bank was seeking to improve the performance of their pre-trade MVA for more timely calculations. This was challenging as they needed to compute their MVA exposures over long time horizons, with a large number of paths. Even with a sensitivity-based approach, this process took close to 10 minutes using a single-threaded CPU calculation. 

Solution: Zanders analyzed the solution and identified several bottlenecks. We developed and optimized a GPU-accelerated solution to ensure efficient GPU utilization, parallelizing the calculations across scenarios and risk factors.  

Performance: Our GPU implementation improved MVA calculation speed by 51x. Improving calculation time from just under 10 minutes to 10 seconds. This significant increase in speed enabled more timely and frequent assessments and decisions on MVA. 

Our Recommendation: A strategic approach to GPU computing implementations 

There are significant benefits to be achieved with the use GPU computing. However, there are some considerations to ensure an effective use of resources: 

We work with firms to develop bespoke solutions to meet their high-performance computing needs. Zanders can help in all aspects of GPU computing implementation, from initial design to the analysis, development and optimization of your GPU computing implementation. 

Conclusion 

GPU computing offers significant improvements in the speed and efficiency of financial calculations, typically boosting calculation speeds by factors of 10-100x. This enables financial institutions to manage their risk more effectively, including the computationally demanding calculations of MVA. By replacing CPU-based calculations with GPU computing, banks can dramatically improve their capacity to process greater volumes of calculations with higher frequency. As financial markets continue to evolve, GPU computing will play an increasingly vital role in their calculation infrastructure.

To find out more on how GPU computing can enhance your institution's risk management processes, please contact Steven van Haren (Director) or Mark Baber (Senior Manager). 

After a prolonged period of stable low (and at points even negative) interest rates, 2022 saw the return of rising rates, prompting Dutch digital bank, Knab, to appoint Zanders to reevaluate and reinforce the bank’s approach to risk.

The evolution of Knab

Founded in 2012 as the first fully digital bank in The Netherlands, Knab offers a suite of online banking products and services to support entrepreneurs both in their business and private needs.

“It's an underserved client group,” says Tom van Zalen, Knab’s Chief Risk Officer. “It's a nice niche as there is a strong need for a bank that really is there for these customers. We want to offer products and services that are really tailored to the specific needs of those entrepreneurs that often don’t fit the standard profile used in the market.”

Over time, the bank’s portfolio has evolved to offer a broad suite of online banking and financial services, including business accounts, mortgages, accounting tools, pensions and insurance. However, it was Knab’s mortgage portfolio that led them to be exposed to heightened interest rate risk. Mortgages with relatively long maturities command a large proportion of Knab’s balance sheet. When interest rates started to rise in 2022, increasing uncertainty in prepayments posed a significant risk to the bank. This emphasized the importance of upgrading their risk models to allow them to quantify the impact of changes in interest rates more accurately.

“With mortgages running for 20 plus years, that brings a certain interest rate risk,” says Tom. “That risk was quite well in control, until in 2022 interest rates started to change a lot. It became clear the risk models we were using needed to evolve and improve to align with the big changes we were observing in the interest rate environment—this was a very big thing we had to solve.”

In addition, in the background at around this time, major changes were happening in the ownership of the bank. This ultimately led to the sale of Knab (as part of Aegon NL) to a.s.r. in October 2022 and then to Bawag in February 2024. Although these transactions were not linked to the project we’re discussing here, they are relevant context as they represent the scale of change the bank was managing throughout this period, which added extra layers of complexity (and urgency) to the project.

A team effort

In 2022, Zanders was appointed by Knab to develop an Interest Rate Risk in the Banking Book (IRRBB) Roadmap that would enable them to navigate the changes in the interest rate environment, ensure regulatory compliance across their product portfolio and generally provide them with more control and clarity over their ALM position.  As a first stage of the project, Zanders worked closely with the Knab team to enhance the measurement of interest rate risk. The next stage of the project was then to develop and implement a new IRRBB strategy to manage and hedge interest rate risk more comprehensively and proactively by optimizing value risk, earnings risk and P&L. 

“The whole model landscape had to be redeveloped and that was a cumbersome and extensive process,” says Tom. “Redevelopment and validation took us seven to eight months. If you compare this to other banks, that sort of execution power is really impressive.”

The swiftness of the execution is the result of the high priority awarded to the project by the bank combined with the expertise of the Zanders team.

Zanders brings a very special combination of experts. Not only are they able to challenge the content and make sure we make the right choices, but they also bring in a market practice view. This combination was critical to the success of the execution of this project.

Tom van Zalen, Knab’s Chief Risk Officer.

Clarity and control

Armed with the new IRRBB infrastructure developed together with Zanders, the bank can now measure and monitor the interest rate risks in their product portfolio (and the impact on their balance sheet) more efficiently and with increased accuracy. This has empowered Knab with more control and clarity on their exposure to interest rate risk, enabling them to put the right measures in place to mitigate and manage risk effectively and compliantly.

“The model upgrade has helped us to reliably measure, monitor and quantify the risks in the balance sheet,” says Tom. “With these new models, the risk that we measure is now a real reflection of the actual risk. This has helped us also to rethink our approach on managing risk.”

The success of the project was qualified by an on-site inspection by the Dutch regulator, De Nederlandsche Bank (DNB), in April 2024. With Zanders supporting them, the Knab team successfully complied with regulatory requirements, and they were also complimented on the quality of their risk organization and management by the on-site inspection team.

Lasting impact

The success of the IRRBB Roadmap and the DNB inspection have really emphasized the extent of changes the project has driven across the bank’s processes. This was more than modelling risk, it was about embedding a more calculated and considered approach to risk management into the workings of the bank.

“It was not just a consultant flying in, doing their work and leaving again, it was really improving the bank,” says Tom. “If we look at where we are now, I really can say that we are in control of the risk, in the sense that we know where it is, we can measure it, we know what we need to do to manage it. And that is, a very nice position to be in.”

For more information on how Zanders can help you enhance your approach to interest rate risk, contact Erik Vijlbrief.

The implementation update covers observations, recommendations and supervisory tools to enhance the assessment of IRRBB risks for institutions and supervisors.¹ Main topics include non-maturing deposit (NMD) behavioral assumptions, complementary dimensions to the SOT NII, the modeling of commercial margins for NMDs in the SOT NII, as well as hedging strategies.  

Some key highlights and takeaways from the results of sample institutions as per Q4 2023: 

• Large dispersion across behavioral assumptions on NMDs is observed. The significant volume of NMDs as part of EU banks’ balance sheets, differences in behavior between customer / product groups and developments in deposit volume distributions, however, underline the need for more solid and aligned modeling. The EBA hence suggests NMD modeling enhancements and recommends (1) banks to consider various risk factors related to the customer, institution and market profile, as well as (2) a supervisory toolkit to monitor parameters / risk factors. Segmentation and peer benchmarking, (reverse) stress testing as well as (combining) expert judgment and historical data are paramount in this regard. The recommendations spark banks to reevaluate forward looking approaches, as shifting deposit dynamics render calibration solely based on historical data insufficient. Establishing a thorough expert judgment governance including backtesting is vital in this respect. Moreover, assessing and substantiating how a bank’s modeling relates to the market is more important than ever. 

• Next to the NII SOT that serves as a metric to flag outlier institutions from an NII perspective, the EBA proposes additional dimensions to be considered by supervisors. These dimensions, which aim to reflect internal NII metrics, must complement the assessment and enhance the understanding of IRRBB exposures and management. The proposed dimensions include (1) market value changes of fair value instruments, (2) interest rate sensitive fees/commissions & overhead costs, and (3) interest rate related embedded losses and gains. It is important to note that it is not intended to introduce new limits or thresholds associated with these dimensions. 

• Given concerns and dispersion regarding the modeling of commercial margins for NMDs in the NII SOT (38% of sample institutions assumed constant commercial margins versus the remainder not applying constant margins), the EBA now provided additional guidance on the expected approach. They recommend institutions to align the assumptions with those in their internal systems, or apply a constant spread over the risk-free rate when not available. Key considerations include the current spread environment, the context of zero or negative interest rates and lags in pass-through. The EBA’s clarification indicates that banks are allowed to apply a non-constant spread. This serves as an opportunity for banks still applying constant ones, as using non-constant spreads enhances the ability to quantify NII risk under an altering interest rate environment. 

• Hedging practices vary significantly across institutions, although hedging instruments (i.e. interest rate swaps) to manage open IRRBB positions are aligned. Hedging strategies have significantly contributed to meeting regulatory requirements, with all institutions meeting the SOT EVE as per Q4 2023, compared to 42% that would not have complied if hedges were disregarded. For the SOT NII, however, 13% of the sample institutions would have been considered outliers if this regulatory measure had been applied in Q4 2023 (versus 21% when disregarding hedges). This result shows that it is key for banks to find a balance between value and earnings stability, and apply hedging strategies accordingly. As compliance with SOTs must be ensured under all circumstances, stressed client behavior and market dynamics must be accounted for. 

In the upcoming years, the EBA will continue monitoring the impact of the IRRBB regulatory package, focusing on NMD modeling, hedging strategies, and potential scope extensions to commercial margin modeling. It will also assess Pillar 3 disclosure practices and track key regulatory elements such as the 5-year cap on NMD repricing maturity and Credit Spread Risk in the Banking Book (CSRBB)-related aspects. Additionally, the EBA will contribute to the International Accounting Standards Board’s (IASB's) Dynamic Risk Management (DRM) project and evaluate the impact of recalibrated shock scenarios from the Basel Committee. 

The EBA publication triggers banks to take action on the four topics outlined above, as well as on hedge accounting (DRM) in the near future. Zanders has extensive relevant experience, and supported on:  

• Numerous NMD topics, including modeling, validation and benchmarking. Furthermore, we published a series of whitepapers regarding NMD modeling concepts and approaches, deposit rate dynamics, forward looking perspectives and migration dynamics in deposits that is particularly relevant following this EBA publication. 

• Drafting an IRRBB strategy, advising on coupon stripping and developing a hedging strategy, thereby carefully balancing value and NII risks (SOT EVE / NII). 

• Validating a hedge accounting framework, developing a hedge account model and hedge accounting outsourcing. Zanders moreover held a survey on DRM as well as published an extensive series of articles on the DRM project of the IASB and its implications for banks. 

Contact Jaap Karelse, Erik Vijlbrief (Netherlands, Belgium and Nordic countries) or Martijn Wycisk (DACH region) for more information.