2020 CCAR Resubmission
Sia Partners supports its clients in the valuation of the impacts in regards to the French health reform (100% Santé) by using the existing models, and by exploiting more complex models with the use of machine learning, notably to enhance predictive accuracy.
The Health sector is regularly subject to regulatory changes. Insurers must initially assess the impacts they can generate from their portfolio, make the best use of the available data, adapt their prices and their product range accordingly, and then finally make sure their contracts are compliant. Sia Partners supports its clients during these three fundamental stages, for all regulatory developments (responsible contracts, ANI, intra-annual cancellations, etc.):
Building a modeling database:
Reading and exploiting data from data centers is an extremely important step in modeling risk and predicting its evolution. This phase is divided into four parts.
1- From management data to a modeling database
Reconciliation of exposure and claims files, database formatting, and data quality control.
2- Understanding data
Improvement of risk knowledge and identification of the specific features of the portfolio available (overall analysis of the portfolio, analysis of data and their influence on the model, etc.).
3- Development and audit of models
Selection and development of models by business constraints.
4- Delivery of the final solution
Delivery of a product that is be perfectly integrated into business processes (assistance with the implementation of the solution, writing of technical specifications, and implementation of data visualization tools for the communication of results).
Our proposed approach to measure the impact of the implementation of 100% Santé and for the compliance is as follow:
The "100% Santé" reform aims to regulate the Health product features causing an increase in number of people canceling or delaying healthcare for financial reasons. Three product lines are then created, the lowest cover includes a commitment to access care and have full reimbursement. Thus, how can the players anticipate the costs generated by behavioral changes of the insured in the portfolio?
We can clarify these impacts for our clients, as well as using machine learning tools that can refine the prospective vision. Traditional pricing techniques have certain limits because they are based on linear links between variables, while certain behavioral effects can be more complex. This approach can be broken down into six stages:
Preliminary phase: Integration into an existing environment
Ensure overall consistency, logic in changes, as well as the respect of strategic, financial, and commercial decisions.
Phase 1: Establish a framework along three main pillars (actuarial, risk, compliance)
Phase 2: Perform preparatory work
Phase 3: Analyze
Phase 4: Make the necessary changes (questioning the technical balance requires a new pricing framework)
Phase 5: Lead operational implementation
Phase 6: Supporting change
The objective of this compliance is to find a technical balance while facing challenges and the new regulatory frameworks.
For this, additional studies can be done, for example, to measure the changes in underwriting behavior caused by the implementation of the reform (for example, to adapt the commercial strategy and optimize acquisition costs).