Esther Use Cases
We offer an analytics framework for pricing and risk management teams based on our compiler software. Our software helps our customers' internal quant and IT teams design, build and deploy new analytics.
Key reasons for investing in the Esther platform include:
Usefulness and quality of models with a fast execution time
Large infrastructure cost savings from replacing a big grid with a single server
Higher modelling productivity Faster time to market
New legal entity-level models provide a better pricing and risk management capability
Large grids cost many $ millions every year, a large server costs $250k (cloud list prices)
Do more with less coding, no explicit performance optimisation required
When should you consider Global Valuation Esther?
Do you need to retain full control of your analytics development and the models you use?
Would you benefit from improved analytics:
Entity-level XVA metrics, including Funding Risk and Cost of Capital?
Wrong-way Risk collateral add-ons and reverse stress testing?
Model risk management?
Esther offers fast, high-quality analytics:
New architecture enables high-quality, top-of-the-house modelling
Ultra fast performance for model execution with large portfolios
Built-in Reverse Stress Testing for actionable strategy insights
Are your objectives to:
Develop analytics in-house with an independent platform?
Complement existing risk systems with new, non-intrusive analytics?
Build a new, next generation pricing and risk platform in-house?
Esther is designed to accelerate in-house modelling:
Fully programmable, high productivity model development and execution platform
Can be integrated with existing systems as loosely or tightly as desired
Esther is the only 2nd generation pricing and risk platform available now.
Partner Companies Offering Solutions
Partner firms can use Esther for building new complementary analytics to their existing software solutions. Global Valuation offers Esther as a development and execution platform and will not build specific end-to-end pricing or risk management solutions ourselves.
We will help end-clients and partners learn how to use the Esther Domain Specific Language (EDSL) for modelling and our platform to solve advanced modelling challenges for large portfolios.
We have developed two example templates, one for Banking and one for Clearing. The templates are free to our clients and partners to use as a starting point for their own implementations. All of their own work will be proprietary to their firms, allowing partner firms build their own software IP independently of Global Valuation. We only charge our clients for the compute engine.
Use Case Examples
As a compiler-based modelling and model execution platform, Esther can be applied to any analytics use case in pricing and risk management. We do not propose Esther as an end-to-end software solution, but a platform enabling the building of such solutions. Our goal is to train our customers' internal teams to use Esther and then step aside.
To facilitate the take-up of Esther, we have developed two example templates:
For Banking we implemented high-quality, entity-level portfolio analytics demonstrating a set of XVA metrics, with Wrong-way Risk and Reverse Stress Testing included in the template model
For Clearing, we created margining analytics comprising Initial and Variation Margin, WWR add-ons, Default Fund contributions and a method of evaluating different margin policies through Reverse Stress Testing
What is Wrong-way Risk and why does it matter?
Wrong-way-risk (WWR) stems from the correlation between market risk and counterparty credit risk
The primary problem is incomplete information: Not having direct access to the full counterparty portfolio, dealers cannot set accurate trading limits to ensure liquidity within the margin period of risk.
Two recent cases that alerted regulators worldwide include the Archegos collapse that ultimately resulted in the failure of Credit Suisse
Another instance was the failure of Tsingshan to post margin that resulted in the unprecedented tear-up of executed but not settled contracts at the London Metal Exchange, leading to litigations with a total possible legal liability of $19.2bn
What are Permissioned Analytics?
Permissioned Analytics is Global Valuation’s solution for secure client-side execution of risk analytics
Thanks to the GV Esther Compiler, risk scripts can execute securely on hardware and data controlled by the data owner.
Credit risk officers at dealers send risk scripts to counterparties.
The counterparty reviews the script, permissions execution and decide whether to submit back the resulting report.
Read our presentation describing how this will work.
Calculate XVA add-ons for pricing incrementally in the context of the entity-level portfolio during a phone call with a customer.
RST based on interactive risk drill-downs on selected extreme but possible loss scenarios, allowing attribution of P&L to specific trade positions, risk factors and counterparties.
Use Esther for complex entity-level models for Funding and Capital management on the full Bank portfolio. Models run fast on a single large server.
FAST INCREMENTAL PRICING
REVERSE STRESS TESTING
Features of our template implementation for Banking:
Our hypothetical scenarios include dynamic credit shocks correlated to market risk factors, thus allowing systematic quantification of WWR risks.
Make model risk management a value-adding activity: Enable model performance monitoring against relevant challenger models.
Esther offers development tools for advanced analytics capabilities, capable of running complex top-of-the-house analytics such as Funding and Capital Valuation Adjustments. Quants can develop models to cover the main analytics use cases across the trade life-cycle in the bank.
Esther enables a leap in Pricing and Risk analytics in banking, when Portfolio Risk Management and XVAs are more critical than ever:
Robust management of Funding Risk escalated by inflation and the Libor transition
Capital and collateral optimisation for increased Capital Markets trading and regulatory compliance
WWR and Reverse Stress Testing capabilities are available with all XVA models
Model Risk to address increasingly complex tail and stress scenarios
Features of our template implementation for Clearing:
REVERSE STRESS TESTING
WWR COLLATERAL ADD-ONS
We generate 10 million hypothetical scenarios for every time step to provide a fine-grained and exhaustive description of loss and default events.
RST is based on interactive risk drill-downs on selected extreme but possible loss scenarios, allowing attribution of P&L to specific counterparties and trade positions.
Our hypothetical scenarios include dynamic credit shocks correlated to market risk factors, allowing systematic quantification of Wrong-way Risk collateral add-ons. Visualise the risk mitigation impact of the add-ons interactively.
The Esther user can define valuation models in full flexibility resting assured that calibration and pricing will be executed fast and accurately.
See our paper a detailed case study on WWR and Reverse Stress Testing
Wrong-way risk and reverse stress testing features are key regulatory requirements Esther can address for clearinghouses. With a large set of hypothetical scenarios (e.g. 10m), analysis of wrong-way risk becomes statistically rigorous. Combining it with reverse stress testing allows clearinghouses to mitigate the risks of arising from sub-optimal margining policies.
Our research shows that new WWR-based policies can dramatically reduce the risk of contagious member failures while retaining overall margin levels at current levels.
Risk Analytics in the Age of AI
The video introduces Esther in the context of Artificial Intelligence and Wrong-way Risk. It covers the technological and mathematical break-through developments that made Esther possible in the past 20 years.
The video also includes a brief demonstration of the Esther modelling language and Cloud IDE.
We come across several primary concerns with many customers: How to improve pricing, risk and capital management at a reasonable cost.
Portfolio Risk Management
Esther was designed from the ground up with entity-level portfolio risk management as a primary use case.
Our template solution for banking demonstrates one way of using new XVA analytics on full bank portfolios, with actionable insights from reverse stress testing.
We can help you design and build a model that works well in your business context.
Speed of Model Execution
Esther’s model execution architecture provides a far superior performance over traditional grid-based engines:
Esther vectorizes calculations, enabling highly efficient use of parallel hardware such as GPUs.
Esther scales vertically first, and only extremely large portfolios with very advanced models require slower horizontal scaling approaches.
Model development productivity:
Esther’s modelling language and IDE offer a high-productivity environment where your developers can focus on business logic rather than the complexities of performance optimization.
The Esther Solver running on a single large GPU server out-performs very large traditional compute grids.