ML-powered SaaS solution automating scenario generation, stress testing, expected credit losses and BS / IS projection
Treasury Pricing & Analytics Enterprise Risk
ML-powered SaaS solution automating scenario generation, stress testing, expected credit losses and BS / IS projection
Overview
Building Blocks
Machine-Learning (ML) powered SaaS solution that automatically generates stress and CECL / IFRS 9 scenarios, projects balance sheet and income statement items on scenarios. The system supplies all necessary data to FusionOptimum for results rendering and report generation.
Regulatory requirements such as CCAR, other stress tests, CECL, IFRS 9 force banks to incorporate macro-economic scenarios in business planning and post-COVID contingency planning. On top of that, Financial Institutions face questions from auditors considering the extreme events we have recently endured.
Flexible
Top down, organization wide view is possible – scope is up to the clients to determine, and there is the possibility for them to use their own models (along with our scenarios). The scenario expansion is automated (from Given Variables) and Support of PPNR (pre-provision net revenue) Calculation
Accurately assess risks
By considering both stable and stressed market conditions, projection of extreme scenarios with probability of such scenarios occurring and inclusion of Macroeconomic variables
Easy and intuitive User Interface
The application allows for rapid adoption in the client organization.
The application generates sophisticated stochastic stress scenario simulations and to derive actionable analytics and recommendations (optimization, hedging). Pre-Integrated to Fusion Risk via FusionFabric.cloud APIs, it delivers the stress scenario data considering latent correlations between the input macroeconomic, market and portfolio specific variables and features – for accurate forecasting and efficient capital management. Two approaches are possible based on if the bank is large or small.
Large banks
1. The core system sends loan and other data to Straterix via API (owned by Straterix).
2. Straterix calculates forecast scenarios that will be integrated into a default probability and loss given default per selected segmentation such as industry.
3. User has access to Straterix UI for configuration and runs
4. Three to five (credit) scenarios per segmentation will be selected according to their risk percentile and send via FusionFabric.cloud API to Fusion Risk.
5. Fusion Risk aggregates and runs ALM – FTP – IFRS module and display the result in different reports / dashboards.
Small banks
1. Straterix has connections to public data source, the bank data is not needed.
2. Straterix calculates forecast scenarios (credit scenarios)
3. User has no access to Straterix UI, the application is used as a black-box service.
4. Scenarios are imported into Fusion Risk via FusionFabric.cloud APIs for ALM/IFRS9
How it looks
Advanced algorithms
Machine learning-based scenario generation for strategic business planning Show less
Jump diffusion
This process is ideal for analysis of tail risk due to its diffusion and shock components. It separates history into stable (diffusion) and shock (jump) periods. Show less
Assess the viability of your trading strategy
Back testing variable expansion using machine learning Show less
Planning tool
Critical for small banks, it uses the views about one given variable and project all other variables and KPIs Show less
Advanced algorithms
Machine learning-based scenario generation for strategic business planning