Chapter 13

Question 1

Discuss how policy and other data can be a source of risk to a life company

Question 2

Discuss how mortality and mobility rates can be a source of risk to a life company

Question 3

Discuss how investment performance can be a source of risk to a life company

Question 4

Discuss how expenses, including the effect of inflation can be a source of risk to a life insurance company

Question 5

Discuss how withdrawals can be a source of risk to a life insurance company

Question 6

List some problems with using past data for assumption setting

Answer 6

• Potential errors
• Changes in daughter according
• Random fluctuations
• Once-off fluctuations and experience
• Too much heterogeneity and groups
• Insufficient data per cell
• Improvements in mortality so adjust for future expected improvements.

Question 7

What are some problems with using the company’s own past data?

Answer 7

• Target market different
• There may be missing data
• The data could be inaccurate
• There may not be enough data for credibility
• Different standards of underwriting
• Too out of date
• Poor raw data, especially old bc of systems
• Small volume
• Changes in mix of homo groups within past data
• Changes in mix of homo groups to which assumptions apply
• Not sufficient at some age ranges
• Extraordinary events

Question 8

List some problems associated with using industry data

Answer 8

• Not enough detail
• Out of date
• Quality depends on quality of data, systems of contributions => could be unreliable
• Since not all organisations contribute, it’s not representative of the whole market
• There are potential errors
• Difficult to adjust to abnormal fluctuations
• Not the same experience due to different target markets underwriting, past investment performance, product and product features
• Difficult or impossible to split the data into groups that match those to whom pricing will apply
• Changes in how data was recorded over time may not be apparent
• Small volume in homogeneous groups or heterogeneous groups.

Question 9

Why could the per-policy expenses of a term assurance contract differ from per-policy expenses for endowment assurance?

Answer 9

• More extensive underwriting for TA => higher initial expenses
• Simpler structure of TA => lower ongoing expenses
• Mix of business by distribution channel may differ
• Commission payable for the contracts might be different
• Inflation occurred and future inflation
• Per-policy expenses affected by relative business volumes
• Companies expenses might change overtime and may be expected to change in future maybe for better efficiency

Question 10

In what situations would using more model points be more appropriate and in which situations would using fewer ones be more appropriate?

Answer 10

More model points:
* More accuracy required
* More available computing power
* Pricing new products
* Calculation of capital requirements
* Complex with profits bonus calculations
* Reinsurance negotiations
Risk: more computationally intensive, not always practical, expensive

Fewer model points:
* High-level strategic planning
* Idea testing of a new product
* Stress testing many scenarios
* Testing sensitivity to assumptions
Risk: incorrect business decisions due to greater scope for errors

Question 11

Why are there more data limitations in the health and care space compared to the life space?

Answer 11

• Smaller policy volumes (CI and LTCI)=> lower credibility of available data
• Lower incidence rates (IP and CI) => lower credibility of available data
• Changes to products and markets over time limits the applicability of past data
• Heterogeneity of products and markets limits the applicability of industry data
• Multistate (various illnesses, stages, severities)
  
  • Subjective definitions (e.g., what counts as “disabled”?)
  • Repeated events (you can get sick many times)
  • Relies on medical records, claims reports, which can vary in quality and completeness.
    Shorter historical series (newer product types, medical advances change trends quickly)
  • Data often fragmented (hospitals, GPs, insurers, government — not all linked)
  • More variation between insurers’ underwriting and claims definitions.
• Morbidity changes more rapidly than mortality (lifestyle, medical advances, policy changes, etc.)

Question 12

Model risk

Answer 12

The model, typically a probability distribution, chosen to represent future mortality, for example, may not be appropriate. This includes the risk that the model points chosen to represent the underlying policy data are inappropriate

Question 13

Parameter risk

Answer 13

The risk that the parameters used with the model may not adequately reflect the future experience of the class of lives insured or to be insured, even though the underlying model may be appropriate

Question 14

Random fluctuations risk

Answer 14

The risk that the actual future experience may not correspond with the model and parameters adopted, even though these adequately reflect the class of lives insured or to be insured