Ch 18: Models (1)

Question 1

Primary objective of modelling:

Answer 1

• Models used to give appropriate advice and to assist in the day-to-day workings of the company.

Question 2

Requirements of a model: (8)

Answer 2

• Model should be valid and sufficiently rigerous for its purpose and adequately documented
• The workings of the model should be wasy to appreciate and communicate. The results should be displayed clearly and the assumptions to be able to use the model documented.
• The model should exhibit sensible joint behaviour of variables
• The output of the model should be capable of independent verification for reasonableness and be communicable to those to whom advice is given
• Model must not be overly complex (results difficult to interpret/communicate) or too time consuming/expensive to run unless required by purpose of model.
• The model should be capable of development and refinement
• A range of methods of implementation should be available to facilitate testing, parameterisation and focus of results.
• The model parameters should:
  * allow for all significant features of the business modelled
  * Inputs should be appropriate to the business and take into account the economic and business environment in which it is operating
  * reflect risk profile of products

Question 3

What is a model point?

Answer 3

• Data record which is an input for a model, could represent a policy or a group of policies, containing data on the most important characteristics
• Data may need to be grouped together into homogenous groups for which each policy in a group is expected to produce similar results.
• It is then sufficient for a single policy in each group to be run and the result scaled up in order to give result for total set of policies in group.

Question 4

Basic features of life insurance models: (5)

Answer 4

• Involves projecting all relevant cashflows
• Cost of setting up supervisory reserves and required solvency margins needs to be allowed for in order to calculate profit flows
• Proper allowance made for guarantees and options - stochastic modelling
• Allow for interactions and correlations between variables (dynamic links) (particularly where assets and liabilities are modelled together)
• Internal time period (frequency) of cahsflow projection must be short enough (monthly) to produce reliable results but not too short that the model is overly time consuming to run.

Question 5

Advantages of stochastic modelling:

Answer 5

• Method allows a probability distribution to be assigned to one or more unknown future parameters
• Expected costs of guarantees and options can be calculated (deterministic approach might produce a zero liability)
• Future parameters can be assumed to vary together as a dynamic set, (particularly useful for modelling with-profit business)
• Output is in the form of a probability distribution from which statistics such as expected value, variances can be obtained

Question 6

Disadvantages of stochastic modelling (2)

Answer 6

• Time and computing constraints (model may have to be simplified, which may lead to unrealistic modelling)
• Results will be sensitive to the assumed values of parameters or distribution of those parameters (subjective)

Question 7

Scenarios where determenistic models may be used:

Answer 7

• Similar results would be obtained as if a full stochastic projection was used
• A quick independent test is required to see if stochastic projection is reasonable