How to use the PODRA Model
Purpose
This text serves as an introduction on how-to use the PillarOne Dynamic Reinsurance Analysis (PODRA) Model with PillarOne.RiskAnalytics (P1RAT).
Prerequisites
It is necessary to have a running installation of PillarOne.RiskAnalytics 0.4.2 or higher. When opening the program the tree view on the left side should contain the PODRA Model (named Podra). For evaluation purposes an external database server, such as MySQL, should be used. As the P1RAT integrated database is native java, it may be only used for testing purposes.
Step by step
First step is opening the application and selecting the Podra Model. Within the Podra Model we wish to start from scratch on. So we use the context menu within PODRA on "Parametrisations" (mouse click right) to create a new default parametrisation. You will find some more parametrisations within Podra/Parametrisations as examples. In case of having a Podra...Parameters.groovy file, you can import it by using the named context menu. Afterwards, name the newly created parametrisation and open it via the context menu of the parametrisation itself.
In the right pane the (empty) model is shown.
The segments to enter the model are Underwriting, Claims Generators, Correlatons, Lines of Business and Reinsurance Program.
Underwriting Information
As a first step the underwriting information should be feeded into the model. Within the Underwriting context menu there is a button add. Name the new Underwriting segment. Within the segment, the data can be edited by doubleclick.
The underwriting information contains: maximum sum insured, average sum insured, premium and number of policies. It is possible to edit serveral risk bands (they will be used for surplus share treaty reinsurance). Add as many Underwriting segments as you like.
Claims Generators
The next step is the editing of Claims Generators. Context menu add creates a new claims generator. The claims generator is more complex. It contains the claims model, the exposure information association, and the underlying underwiting information.
For the claims generator the type of the claims generator can be selected: Attritional, Frequency Severity, Frequency Average Severity or External Severity. Depending on the selection it is possible to define the appropriate claims generator. The model starts with drawing a the selected claims generator. When the use of a surplus share treaty is expected, the second step is the allocation of an exposure (sum insured, pml, eml) to the drawn loss. Therefore the option to selecting the exposure information allocation can be used.
The combo box "claims generator is based on" is used to scale the random distributions by underwriting information the underwriting segments.
Correlations
Claims generators may interact / correlate. Correlations can be introduced by adding the Correlations Matrix using the Context Menu Add of Correlations on the right pane.
Lines of Business
Now the risks are defined and may be combined or allocated on business segments. These are defined in the Line of Business area. Here again, it is possible to add another Line of Business. The next step is to add different Claims generators (or shares of them) to the dedicated Line of Business. Additionally, to each Line of Business the Underwriting Information is attached, so we can select this additionally.
Reinsurance Program
Individual Reinsurance covers can be added to the Reinsurance Program section. The reinsurance cover consists of a contract (including the type), the inuring priority, the covered lines of business, as well as the covered claims generators. The covered lines of business and the convered claims generators are sent to the reinsurance contract to be covered.
The inuring priority provides the information when an specific contract is used and from which level e.g. the gnpi is taken. For the reinsurance program itself the type has to be specified. Depending on the type there are several options for the necessary parameters.
Result descriptor
Usually, not all possible result variables are generated during a simulation run. This is for performance and disk space reasons. Within the PODRA Model (tree view) several result descriptors can be defined. By opening a result descriptor in the right pane a tree view with selectable output variables is shown. The selected output variables will be found in the Results after a simulation has been done.
Evaluation / Simulation
On Parameters and Result Descriptors within the left tree view it is possible to start simulations using the context menu. The Parameter set and the Result Descriptor to be used have to be selected in the Simulation window on the right pane. After adding the number of simulatons the Results of the PODRA Model contain an additional result set.
Open the result set to see which variables have been generated. On the result variables different risk measures (exactly different functionals) can be added by pressing the appropriate buttons. The full set can be copied to the clipboard by selecting the top left element and pressing ctrl-c.
Further information
