Modelling & Data

Question 1

State the biggest influences on data quality from different companies (3)

Answer 1

1. age of company
2. size of company
3. age of computer and information system
4. existence of legacy systems
5. integrity of systems, e.g. automated checks
6. quality of managers
7. quality and training of staff
8. nature of organisation, e.g. direct insurer vs reinsurer
9. sales channel
10. classes of insurance sold
11. type of insured, e.g. large commercial risks vs smaller personal risks
12. product design
13. rate of change in products
14. simplicity and ease of use of computer systems

Question 2

List the main reasons data may be inadequate and hence lead to uncertainty in modelling (11)

Answer 2

Poor data due to:
1. inadequate data recording procedures
2. poor system design - not capturing all rating factors
3. poor system design - not retaining a history of risks
Inconsistent or incorrect data due to problems in recording:
1. no validation/reasonableness checks on digits, dates, amounts etc
2. no spot checks
3. no recon
4. no cross checks between policy and claims info
5. no cross checks over different policy years
Incomplete (fields missing) due to:
1. poor systems design
2. poor user processes
3. inadequate information collected at outset by the insurer
4. inadequate information collected at the outset by third parties
Non - existent data for:
1. new classes of business
2. new territories
3. unusual risks
4. infrequent claim events
5. certain rating groups
6. extreme values
Data not completely relevant due to changes in:
1. external factors
2. policy terms and conditions
3. underwriting philosophy
4. claims management processes

Question 3

Outline the factors that should be considered when establishing a central computer system for private motor (10)

Answer 3

Function of the system, input from all users will be required:
1) administration
2) accounting
3) actuarial/statistical
4) management
5) marketing
6) other (reinsurance, underwriters, claims, investment, computing)
For private motor there is potentially lots of data to be processed and all users will rely heavily on any system, input is therefore important

Structure or outline of the system:

1) how should the system work
2) will all branches and agents have direct access to the system?
3) for private motor the company will likely want to provide online quoting
4) What type of computer or software will be used?

Cost of implementation

Security:

1) how will the system be monitored
2) who will have access to the system?
3) data protection laws should be considered, particularly if the system is to be used by third parties

Output requirements from the system:

1) premium quotations (possibly linked to broker systems)
2) policy documents
3) renewal and endorsement notices

Question 4

List the data you would expect to see on an individual policy file for private motor (8)

Answer 4

ADMIN DATA
1) Policy number : to define the policy uniquely and provide a link to the related file
2) Name and address of the policyholder for correspondence
3) Broker/branch/comm : for accounting use (also management & marketing info)
4) Dates on risk : defines the exposure period and when a renewal notice is needed
5) Original year of issue : for statistical and management info
6) Endorsement data : indicates a change to the policy; prior data needed so that claim analyses are accurate
7) Premiums : needed for reserving, pricing and monitoring
8) Commission : to check that correct payments are made
RATING DATA
1) Cover : Comp, third party fire theft etc
2) Type of use : SDP, business or domestic
3) Details of driver : name, dob, occupation, gender
4) NCD
5) Excess : compulsory or voluntary
6) Details on car : make, model, engine capacity, parking, age etc
Claims data should then be linked to policy data with a file reference