Data Flashcards
(38 cards)
Sources of data
- Benefit providers own data
- Population data
- Reinsurers data
- Market data
- Other sources - actuarial consultants , overseas
Data considerations
Relevance : similar class of lives and policy conditions to those expected for the benefit
Credibility : implies a sufficiently large volume of data to enable precise estimates to be calculated
Cost of obtaining and analysis no the data and suitability of format
When using past data to set assumptions - insider how to deal with:
- abnormal fluctuations
- changes of experiences with time
- random fluctuations
- changes in the way in which the data was recorded
- potential errors in the data
- changes in the balance of any homogenous groups underlying the data ( business mix)
- geterogeniry within the group to whom the assumptions are to relate
Mortality and morbidity data
- medical advances
Different depending on the source and the underlying group/populaton - differs by type of the employement or social class
Future projections
- claim incidence rates , claim termination rates, average claim size
- mortality and withdrawal rates
-yields to maturity on real and nominal government bonds
-inflation risk premiums - estimates of future risk premiums on other assets
- salary and pension increases
- contributions or premiums
-expense
PMI - Average claim size :
Dep on
- inflation , current protocols and hospital charging structures
Availability of treatments under state healthcare
Mortality projection methods - two deterministic ways
1) Process-based projections - model trends in causes of death.
Drawbacks: death classification and insufficient understanding of the major cause of death processes
And lack of understanding of relationships between diseases
- can include scenario testing
2) Extrapolative methods- trends in mortality are projected in the future.
Drawbacks- subjectivity associated with the choice of period over which such trends are to be determined.
Mortality improvements - stochastic model or simple assumptions
Pros and cons of using population data
- important source of information for health and retirement where benefits depend on mortality and longevity and very few insures and benefit providers have enough data for credible actuarial investigations .
- readily available
- free
- size of the data available makes it credible
Drawbacks:
- the accuracy and reliability may be questionable
- especially in healthcare where definitions may be subjective and the scope for double counting is significant
- may not be available in the appropriate format
- the population experience may be somewhat out of date before it becomes published
- the national experience may not be relevant to a benefit provider;s subset of lives
Since insured and uninsured lives are likely to behave differently - particularly evident in the increased utilisation of medical services in terms of frequency and cost of treatment when people are insured
Reinsurer’s data
- provide statistics relevant to risk to be insured - assist in pricing and reserving
- the statistics are accurate as possible
- reinsurers will have access to more data that is diverse since they deal with many insurers
- charge by reinsurer for providing this service - but provided on the basis of accepting a share of profits from the business written
- actuary must assess the divergence of reinusrer’s data from their compnay’s future expernece and make appropriate adjustments - adjust for their compna’s specific underwriting and claims management
Market data
Insured lives data - industry data
- agreement between benefit providers to pool claims and policy statistics so that overall industry-wide information may be available that would be more credible - regulatory facilitated
- can be used as a base reference for experience monitoring purposes and for setting assumptions
- it represents insured experience while is more relevant
- generally of sufficient volume to to be statistically credible
- reflects local benefit providers - reflecting market practices
- drawback : average market experience and not strictly relevant t any one benefit provider
- actuary will need to consider the extent to which a benefit provider’s own experience will differ from this market average and make adjustments
Published returns
* statistics of this type can also be drawn from the unlisted returns to the local regulator
- benefit providers may end to publish returns to their local regular to at regular intervals - who then pubslishes an annual industry report which freely available on its website
- may be more detailed - number or policies in force
- this information can provide a high-level check on premiums produced from other sources
Other sources
- overseas :
Less relevant than equivalent population an insured lives data of the home territory - even if it is insured, carefu with differing cultures differing state healthcare provision , differential market practices , differing legislation and differing policy conditions
Actuarial consultant
- charge a fee for their services
Data validation
-
recording accuracy
Important for claims and policy data is accurate -
controls on data acceptance
( the software for accepting data inputs, should have built checks that prevent erroneous items from being accepted ) ( certain errors may have exceptions that can only be overwritten by a person of a pre-specified status) -
compulsory fields
( an individual policy record will have certain fields that are mandatory - the input will not be accepted unless all such information is included ) ( age , gender and benefit ) - regular vetting and spot checks
- auditing the accuracy of information will entail the regular inspection of the process by which data are accepted by the system as well ass considering whether the data captured are comprehensive
- policy or member records should be checked “E2E” - from input to eventual use
- Member data and data checks
Member data and data checks
- A reconciliation of the total number of members and changes in membership , using previous data and accounts
- Checks for existence of new members - can be validated against employer, payroll or insurance data to ensure that there are no omissions
- Comparison of average benefits level or of average values of components of the benefit calculation
- Checks for consistency between salary-related contributions and in-payment benefit levels indicated by membership data and the corresponding figures in the accounts
- Minimum and maximum levels of benefits , their components n , ages - checking the individual items of data on each member’s record, a comparison can be made between previous and current records for members who have been active throughout.
Asset data
- to place a credit value - obtain a full listing of the individual assets held
- these individual holdings should then be checked o detrmine whether they are permitted or ae subject to valuation restrictions imposed by scheme rules or legislations
- a retirement scheme my also have assets in the form of insurance policies - details of these policies will be required - including any policies covering death-in-serviece benefits
For funded benefit schemes - the confidence in the asset data may be increased by cross-referencing against -
* audited accounts - this is signed off by a practising auditing professional
* investment manager reports - giving listing of assets held, purchased and sales
*consistency between investment income and assets actually held
Such document are not just useful for valuation but to check other aspects of scheme operations
- asset allocation - to check that assets held are appropriate to the liabilities in accordance with any professional guidance for th actuary
- self-investment - check compliance with any restrictions on any investments related to the1 sponsor
General assumption setting consideration
- Use of the assumption will be
- Choice of the assumption that will have the most financial significance
- potential financial significance of errors in the assumptions will also determine the degree of prudence - Achieve consistency between various assumptions
- Consider any legislative or regulatory constraints
- Consider the needs of the covered lives
- Ensure that the parameters serviced from data are produced as accurately as the body of data will permit
- Ensure tha the data used to derive these assumptions are relevant to the risks that the policies encompass
- Ensure that bases used for periodic valuation and reserves are flexible to reflect changing risk circumstances
- Sensitivity test, especially those assumptions which are most significant in determinin given outcome
General assumption setting process
- Investigate historical experience and make best estimates of the parameters from that experience
- Consider what the conditions , including commercial and economic environment, will be like in the future period for which the assumptions are being made
- Determine what the best estimates of the assumptions will be , given the expected future conditions
- The best-estimate may need to be adjusted in order to include a margin for prudence
The demographic assumption - the expected future experience will depend:
- the target market for the benefit
- the underwriting controls applied or not applied at the start of the policy
- the claim controls applied , or not applied at the claims stage
- the benefit provision wording
- the expected change in the xperience since the time of the last historical investigation to the point in time at which the assumption will on average apply
Claim incidence rates assumption calc
- reinsurance or makret data may come in the form of incidence rates
- otherwise these have to be calculated using exposed to risk techniques
The rates will be broken into discrete risk cells
The split for LTCI:
- age
-Gender - smoker status
- occupation -
Size of benefits selected - type of benefits selected
- class of product
- individual or group
- distribution source
- region
- underwriting status
- duration since policy entry
- after rates for each disease or whatever is calculated
- project it to be future experience by each cause in each future calendar year using a projection factor
- the projection factor could also be age specific as the changes in morbidity over time may not be the same at every age
Trends that may impact the incidence rates -
Social and economic welfare
1. Changing attitudes to health , greater access to healthcare services, better diagnosis of medical ,conditions, and the avilability of new technology
2. The future plans for state welfare provision
3.The economic well-being of the country
4. Inflation generally
Changing health environment
Government promises in welfare provision
- the more generous the state promises in relation to long term care support and medical acre provision - the more difficult it wi; be to acquire and retain business
Pessimism in the economic and the level of unecomment
- this may encourage employees to take advantage of group PMI cover while they are still employed
High inflation
- this will impact the renewal price of PMI and is likely to reduce renewal rates
Recovery rates
- where the benefit will cease if the policyholder recovers from a disability or dies - is usually unlikely and therefore is ignored in pricing
Mortality rates
Non-claiming vs claiming policyholders
( healthy or active ) vs claiming or pensioners or deferred
- mortality of active or heathy - likely to be standard mortality table with adjustments for future
- of claiming ( LTCI - sick ) - the mortality assumption would merit much more effort in order to estimate the future rates as mortality of sick lives is a lot less predictable than he population
Claims within the survival period
- special consideration needs to be given to survivorship requirement for CI
- this is allowed for by an adjustment to the claim incidence rate
Policies without a death benefit
- - shouldn’t’t be over estimated as it would lea to contracts being under costed
PMI lapse rates dep
- is there a better competitor product
-rate of premium or contribution increase - the perceived service delivery by state
( waiting lists etc ) - the current confidence in the economy
( where PMI competes with state )
-sales channel can have a direct effect on persistency or renewal , experience
( high persistency can be associated with insurance intermediaries ) - after-sale service ( mostly affected short term where there is a regular contract)
Persistency rates are likely to be affected: - the level of financial sophistication
Whether it was the policyholder who initiatives the sale
The effect of lapses
- an early lapse can cause financial loss - where insufficient premiums or contributions have been received to recover the intial strain , that is commission and initial expense , involved in the writing the policy
- if no surrender is payable and +ve asset share has been accumulated - the lapse can benefit he insurer
- the effect of the lapses may be selective - the healthier lives will be more likely to lapse , leaving a worsening propensities t claim amongst those continuing
- ## lapse and re-entry
Economic / financial assumptions
- benefit amount
- benefit inflation
- expenses
- expense inflation
- commission and clawback
- investment return
- taxes
- price inflation
-pension increases - earnings inflation
Alternative reimbursements - result in some transfer of risk relating to the cost of providing the services to the healthcare provider
- indemnity or fee for service
- modified fee for service
- per-diem
- per case
-Capitation - salary