MPI Model is a comprehensive reinsurance pricing and risk management system for property-casualty reinsurance companies. The system allows a reinsurance company to monitor the risk profile of their existing portfolio, forecast cash flow, and optimize cash reserving and investment strategy.
The system's analytical engine is based on proprietary Monte Carlo modeling techniques and Bayesian statistical models. The system allows reinsurance company staff to perform the following tasks:
- Enter and maintain submission information in a company-wide database
- Perform flexible and consistent pricing of contracts based on submission information
- Accumulate loss experience from existing accounts
- Monitor and visualize loss development in existing accounts
- Create projections (forecasts) of the loss development process and cash flows
- Compare new information with projections and to revise projections as time progresses
- Perform company-wide reserving and asset-liability matching
MPI Model employs contemporary actuarial science concepts and combines with them advanced software application design features. Among them:
- Fast Multi-variable Frequency/Severity Simulation Algorithms
- Quadratic Optimization
- Discrete Bayesian Analysis
- Object-oriented Database Design
- Flexible Object-Oriented Modular Structure
- Interactive Visual Analysis
MPI Model was designed to price/reserve treaties (policies) with highly non-linear conditions, such as Specific and Aggregate Limits, Dividends or Commutations. That requirement posed problems for conventional analytics, no matter how sophisticated they were. Thus, MPI Model employs Monte Carlo Simulations, which we believe is the only way to effectively price and reserve sophisticated real-life contracts.
The simulations framework serves as a foundation for numerous unique features of MPI Model:
- MPI Model offers a high level of customization, specifically, the ability to define new lines of business and to employ proprietary models
- MPI Model saves statistical assumptions used for pricing for future reuse in reserving
- Using Bayesian Analysis Model updates statistical distributions as new evaluations of data become available
- Integrated pricing/reserving design allows MPI Model to price Loss Portfolio Transfers and Stop Loss treaties
- The ability to price/reserve highly non-linear contracts allows MPI Model to analyze single policies as well as portfolios
- Simulating the Loss Distribution for every policy, MPI Model is capable of generating a company-wide Aggregate Loss Distribution and of estimating necessary Bulk Reserves
- Simulations allows MPI Model to produce not only final Ultimate loss estimates, but to predict payout for any given point in time, which supports proper Discounting of Reserves and IRR calculations
- Proper discounting provides a solid basis for Profitability calculations for every single contract as well as any market segment or any other subset of the full portfolio
The Object-oriented Database design featured in MPI Model guarantees
- High Data Integrity
- Effortless Renewal processing
- Consistent Versioning and Audit Trail
The application's modular design allows MPI's development team to endlessly modify, improve and customize the software to meet a customer's demands as well as reflect the changing nature of the Insurance/Reinsurance business.
In short, MPI Model is a must have tool for any company with a significant portfolio of long-tail LOBs.
MPI Model is implemented as a client/server application utilizing Oracle 8 Database Sever as a backend with client workstations operating in the Microsoft Windows environment. Data integrity and consistency is supported by numerous stored procedures and triggers. Extensive data importing and exporting capabilities are implemented through OLE Automation with Microsoft Excel.