CH38 - Monitoring Flashcards
Reasons for monitoring experience (3)
Monitoring the experience is a fundamental part of the actuarial control cycle. The experience will be monitored so as to:
- update the methods and assumptions so that they are more relevant to future experience
- monitor any trends in experience, particularly adverse trends, so as to take corrective actions
- provide management (and other key stakeholder) information
Data required for monitoring
The basic requirement is that there is a reasonable volume of stable, consistent data, from which future experience and trends can be deduced.
Consistent here means that, when comparing the experience of one group with another, the data used as a basis for the calculations for each group should be:
* in a similar form
* preferably extracted from the same source
* grouped according to the same criteria
* equal in terms of reliability
The data ideally needs to be divided into sufficiently homogeneous risk groups, according to the relevant risk factors.
However, this ideal has to be balanced against the danger of creating data cells that have too little data in them to be credible.
As well as data on the feature being assessed, it is necessary to have data on the exposure to risk, divided into the same cell structure as the experience data.
Analysis process - Statistical factors
Once the data have been grouped in an appropriate manner, the analysis can be performed.
For statistical factors, such as mortality, morbidity and withdrawal, this will involve the calculation for each age band of the number of deaths (or withdrawals) divided by the number exposed to risk of death (or withdrawal).
The results can then be compared with assumptions or standard tables (if these exist)
Use of the results
The results of an analysis of experience should not be used blindly. Consideration should be given to whether the period under investigation was typical and whether the experience is likely to be representative of future experience.
Allowance should be made for:
- trends (including gradual changes over time)
- cycles
- abnormal events
- random fluctuations
Before being used as assumptions, the results of experience analysis should also be adjusted to allow for data and modelling risk.
Monitoring and control cycles
Monitoring of experience is fundamental to effective implementation of the actuarial control cycle and the risk management control cycle.
This is an iterative process as it may result in:
1. changes in assumptions or models used, e.g. in pricing, setting contributions, and provisioning
2. a change in the assessment of the risks faced by a provider or in its risk management strategy.
List at least ten tasks and/or investigations that a life insurance company performs that require assumptions about future experience
- Pricing
- Provisioning - statutory assessment
- Provisioning - realistic assessment
- Profitability monitoring
- Modelling capital requirements
- Investigating the resilience of the company to adverse future experience
- Determining appropriate types and amounts of reinsurance
- Determining appropriate underwriting policy
- Investment strategy investigations, e.g. asset-liability modelling
- Analysis of surplus investigations
- Setting discontinuance terms
Analysis process - Economic factors
Economic factors usually have the greatest impact on the result for a company or scheme, but are also generally outside the management’s control.
Economic factors include:
- interest rates and investment returns
- expense inflation
- salary growth
For interest rates and investment returns(which are the main economic factors for a benefits scheme or insurance company), the analysis is simply a comparison between the actual returns and those assumed.
For expense inflation, the actual level of expense inflation would need to be determined by removing from the expense analysis any costs that were included only in the previous or current data, and by considering the change in unit costs rather than overall totals (in order to remove the impact on total expenses from volume changes)
A salary growth analysis is likely to distinguish between general inflation-related salary increases and promotional salary increase.
Explain how the data may be subdivided when analysing expenses
The main items of expense for a financial provider are:
- salaries and salary-related expenses
- property costs (rent, property taxes, heating, lighting and cleaning)
- computer costs
- investment costs (investment department, stamp duty, commission, custodian etc)
In order to analyse these expenses, the provider will look to subdivide them into:
- fixed / variable
- direct / indirect
- initial, renewal, termination and investment expenses
- product with which they are associated
- whether they are proportional to the:
- number of contracts
- amount of the claim (or benefits)
- amount of the premiums (or contributions)
- amount of funds under management
The reason for these subdivisions is that they correspond to how assumptions are made in most actuarial models
In translating the results of an experience investigation into assumptions, the purpose for which the assumptions will be used is an important factor.
State the other factors that should be taken into account (7)
Other factors to take into account in setting assumptions are the:
- significance of particular assumption to the overall result
- relationships / consistency between the assumptions
- the needs of the client
- any legislative or regulatory constraints
- margins in assumptions vs risk discount rate
- time horizons over which assumptions will apply
- credibility of results and hence need for margins
Monitoring process - exam type question issues
Monitoring investigations typically involve the following stages:
- the division of data into suitable groups that are homogeneous by risk - need to consider:
- the volume of data in each cell (its credibility)
- the risk factors for the investigation (e.g. age, gender)
- changes that have occurred that will reduce the relevance of old data
- identification of any past trends, cycles and anomalies and random variation in the past data
- using the results to revise models and assumptions used - need to consider:
- the purpose, and hence the need for accuracy and margins for prudence
- allowance for future trends
- likely differences in future experience from past experience
These are the types of issue that would need to be considered if an exam question is asked about an analysis. This list may be particularly useful if the analysis is unusual.