Prophet Model Enhancement

Project Background

With the full implementation of IFRS 17 and next-generation capital regimes such as ICS, actuarial models have evolved from simple calculation engines into core systems that underpin financial reporting, capital management, and strategic decision-making. As product structures and modeling logic grow increasingly complex, legacy Prophet liability models often face challenges such as performance bottlenecks, excessive structural complexity, high maintenance costs, and rising model governance risks.

Under the pressure of high-frequency month-end closes, scenario analyses, and audit reviews, model stability, explainability, and sustainability have become critical determinants of operational efficiency and risk management quality. This project was initiated in response to these challenges, with the objective not merely to “make the model run faster,” but to systematically enhance the model’s long-term value through comprehensive structural and governance improvements.

Project Objectives

This project was designed with clear, practical objectives aligned with actuarial and operational needs:

• Significantly reduce runtime and re-run cycles for IFRS 17 liability models
• Simplify model architecture to improve transparency and interpretability
• Strengthen model governance and auditability while reducing key-person risk
• Establish a robust foundation for future automation and model integration

Solution Highlights

1. Systematic Performance Optimization
A comprehensive review of the model structure and calculation flow was conducted to remove unused or inefficient variables and formulas, consolidate duplicated product logic, and optimize table access and execution logic within Prophet. These enhancements significantly improved runtime performance without compromising calculation accuracy.

2. Structural Simplification and Transparency
Complex and hard-to-trace extended formulas were converted into standard variables, and clear indicators were introduced to distinguish different calculation logic and scenarios. This approach greatly improved model readability, validation efficiency, and communication with stakeholders and auditors.

3. Enhanced Model Governance and Maintainability
By consolidating libraries, restructuring master products, and standardizing naming conventions, the project established a consistent and extensible model architecture. This reduced ongoing maintenance effort, improved handover capability, and supported more sustainable team-based model management.

4. Readiness for Automation and Future Needs
The model enhancement was executed with future scalability in mind. Output variables were streamlined to reduce downstream data-processing loads, DCS and data interfaces were reviewed and optimized, and flexibility was preserved for integration with workflow automation platforms and near-real-time analytics.

Project Outcomes

The enhancement delivered measurable and impactful results:

• Over 80% reduction in total runtime for non-stochastic IFRS 17 liability models
• Significantly faster re-runs and testing cycles, supporting tighter month-end close timelines
• Improved model clarity, explainability, and audit readiness
• Reduced maintenance effort and operational risk, enhancing long-term model sustainability

Project Value

The true value of this project extends beyond performance gains. It successfully transformed the actuarial model from a high-maintenance, opaque calculation engine into a stable, transparent, and well-governed core platform.

By strengthening the model foundation, actuarial teams can now redirect effort from troubleshooting and manual intervention toward analysis, interpretation, and decision support. This positions the organization to respond more effectively to evolving regulatory requirements and business needs, while building a resilient actuarial infrastructure capable of sustainable long-term growth.