Modelling Flashcards
Producing a solution
Commercially produced product
Modified existing model
New model
Construction of actuarial model
Any model should be fit for the purpose for which it is being used
Operational issues to be considered include:
• be well documented
• be easily communicable, with clearly displayed results
• have sensible joint behavior of variables
• be capable of independent verification for reasonableness
• not be overly complex or time-consuming to run
• be capable of development and refinement
• be capable of being implemented in a range of ways
• have an appropriate time period between projected cashflows, balancing the reliability of the output with the speed of running the model
Developing a model
• specify the purpose and key features of the model
• obtain and adjust the data
• set parameters/assumptions, including any dynamic links
• construct the model cashflows
• check the accuracy and fit the model, and amend if necessary
• run the model as many times as required
• output and summarize the results
Use of models for pricing
Model point: represent relatively homogenous underlying groups of policies
Risk discount rate is used to discount the future net cashflows, it allows for:
• the return required by the company
• the level of statistical risk (assessed analytically, by sensitivity analysis, from a stochastic model or by comparison with market data)
A stochastic risk discount rate could be used as well
Use of models for setting future financing strategies
Used by benefit schemes to determine future financing strategies
Results of model give the amount and timing of future contributions
It is acceptable for a scheme to have a deficit as long as the sponsor can make up for it and there are sufficient assets to meet benefit outgo as it falls due
Use of models for risk management
Can be used to determine capital requirements
Use of models for assessing provisions
The valuation of a company’s liabilities for regulatory purposes is likely to be carried out in each individual policy or member, rather than by using model points
Use of models for pricing options and guarantees
Option and guarantees are likely to be priced by using a stochastic model, particularly if linked to an investment outcome
Model and parameter error
The results of the model depend on the model itself and the values assigned to parameters
Sensitivity analysis is used to illustrate the potential variability of the results and to identify the impact of mis-estimation of the parameter values (varying individual assumptions and assessing the impact on the results)
Scenario testing involves changing many assumptions in combination
Goodness of fit tests help to reduce model error
Alternative ways of allowing for risk
Statistical risk associated with parameter values can be allowed for in the discount rate and/or by including margins in the parameter values