Chapter 12.2 Assumptions (1) Demographic assumptions

Question 1

Briefly describe a common framework which can be used to drerive assumptions (5)

(demographic assumptions)

Answer 1

1. Investigate past experience; make past best estimate parameters; appropriate in context of historical conditions/then-circumstances
2. Consider future conditions (including commercial and economic environment ) during period for which assumptions will be used
3. Determine future best estimates assumptions, given expected future conditions
4. Extent of (a) relying past data vs (b) allowing for other factors, depends on data credibility/relevance + parameter’s predictability
5. Adjust best estimates with margin. Size of margin depends on:
   
   • purpose for which model is required
   • degree of risk associated with parameter

Question 2

Main demographic assumptions

List some of the main demographic assumptions for health care contracts (5)

Answer 2

• PMI Incidence Rates
• Critical Illness Incidence Rates
• Long-Term Care Transfer Probabilities (incidence and benefit amount)
• Other claim inception rates
• Persistency rates are also important for health care insurance conracts

Question 3

Key steps within framework for deriving demographic assumptions:

Briefly list key steps to be taken when deriving demographic assumptions, given the common framework used to set assumptions (5)

Answer 3

• collect appropriate data
• grouping data
• adjusting the data
• calculate the rates
• adjusting the rates

Question 4

List key requirements to be satisfied for the data used when setting assumptions (7)

Answer 4

Key data requirements: data should be

• relevant (similar class of lives to that being insured)
• credible (sufficient volume to draw conclusions/remove random flactuations)
• reliable
• readily available

Consideration should also be given to

• the cost and suitability of the data given the purpose for which the eventual assumptions will be used
• the extent of need to adjust data for credibility and relevance
  *

Question 5

Collecting appropriate data for assumptions

What conflicting aspects exist between the period of time with respect to which data is collected*, and the *volume of data used for setting assumptions? (2)

Answer 5

2 important points regarding data used for assumptions

• data would relate to an appropriate period of year, such that volume is adequate..
• …but excessive heterogeneity due to changes over time is not introduced

Question 6

List various data sources which can be used for setting demographic assumptions (5)

Answer 6

Data sources

• Own data
• Population data
• Reinsurer data
• Market data
• Other sources

Question 7

Collecting data: data sources, own data

Briefly touch on the use of the insurer’s own data for setting assumptions in terms of

• relevance (4)
• how own data ranks compared to other data sources (3)
• volumes of data (2)
• limitations of using own data (2)

Answer 7

Own data is highly/most relevant in terms of

• underwriting approach,
• policy condition,
• claims management and
• distribution channel

Own experience is best case.

• cost, reliability and format are not issues

Volume

• may not be sufficient for credibility, particularly for risk cells.
• may need adjustment

Limitations

• for brand new entrants into market, there will be limited own data, and so such insurers will have to be totally reliant on external sources

Question 8

Collecting data: data sources, population data

Briefly touch on the use of population data for setting assumptions in terms of

• where the data may come from and a key point related to population data (1)
• advantages of using population data (4)
• disadvantages of using population data (6)

Answer 8

How population data may arise and a key point related to it:

• government may periodically produce analysis of healthcare experience for population
• key point related to population data: for some risk cells, it may be the data available to form a starting point for assumption setting

Advantages:

• readily available,
• often free.
• usually high volume and credible

Disadvantages:

• Nation data may not be relevant to insured lives. Better if data is split by groups, regions, age
• Insured population may behave/utilise differently to national population (i.e. utilise more)
• Sickness and incapacity definitions will be different to ‘claim’ definitions in policy
• Accuracy and reliability may be questionable, especially where definitions are subjective. Scope for double counting is enormous
• Data may not be available electronically, or in an appropriate format
• May be out of date before it is even published

Question 9

Collecting data: data sources, reinsurer data

Briefly touch on the use of reinsurer data for setting assumptions in terms of

• how the insurer would gain access to reinsurer data (1)
• advantages of using reinsurer data (3)
• disadvantages of using reinsurer data (5)

Answer 9

How does insurer gain access to reinsurer data?

• Reinsurer may have data as part of its services offered to insurer

Advantages

• There is motive for reinsurer data to be as accurate/fair as possible since profits under reinsurance treaty depends on calculated premiums
• Reinsurers draw information from other companies, and other countries. Excellent for launching new product lines. Will usually provide data as part of a contract
• Credibility will vary, but more credible than insurer’s own data

Drawbacks/important points

• Will need to consider divergence of own future experience from reinsurer data and adjust accordingly
• Differences could arise due to differing claims definitions, claims management processes, underwriting processes, etc
• The use of reinsurance comes at a cost;
  
  • eg reinsurer will likely provide data in return for share of insurer’s business
  • this will need to be factored into the choice of using reinsurer data and whether is is worthwhile

Question 10

Collecting data: data sources, market data - intro

What kind of market data may an insurer use to set assumptions? (2)

Answer 10

• insured lives data (industry data)
• returns to an insurer supervisor/regulator

Question 11

Collecting data: data sources, market data - insured lives/industry data

Briefly touch on the use of the market data (insured lives data) for setting assumptions in terms of

• how this data may arise/be obtained (2)
• in terms of credibility and relevance, how does this data source compare to insurer’s own data and population data? (2)
• advantages of using this data (6)
• drawbacks of using this data (4)

Answer 11

Insured lives data (aka industry data)

• may come/be obtained as a result of an agreement between insurers to pool claims and policy statistics for industry-wide collection. E.g. CMI in UK
• sometimes form the basis of mortality tables

Biggest considerations/issues:

• more credible than insurer’s own data, but less relevant.
• more relevant than population, but less credible

Advantages:

• usually compiled by experts
• highly credible and may be of sufficient volume
• represents insured experience (as opposed to, eg national stats, which include experience of non-insured lives)
• should reflect local companies and market practices within which insurer likely operates…
• …benefits all, as individual companies may get an analysis/comparison of their own experience compared to market
• national relevance

Drawbacks:

• relvance questionable, as only represents a market average, not relevant to any one company
  
  • policy conditions, underwriting and claims management may differ
• may take years for sufficient data to accumulate, depending on level of participation of various market participants
• industry wide data may be very limited; usually available for long term business classes (LTCI, CI), but not for short term (PMI)

Question 12

Collecting data: data sources, market data - returns to an insurance supervisor

Briefly touch on the use of the market data (insured lives data) for setting assumptions in terms of

• how this data may arise (1)
• briefly list advantages of using this data (3)
• brielfy list disadvantages of using this data (2)

Answer 12

Source of data

• some territories require insurers to submit returns to supervisory bodies/regulators

Advantages of using this data

• Data may be credible
• Data may be cheap
• Can use to check position of premium rates compare to others in the market

Diadvantages of using this data

• Relevance questionable, as returns usually used to assess solvency, not price
• Although possibly credible/cheap, formatting/degree of detail are often issues

Question 13

Collecting data: data sources, other sources - intro

List other sources of data an insurer may use for setting assumptions (4)

Answer 13

• Trade Magazines
• Actuarial Consultants
• Overseas Data
• Rate Table Software

Question 14

Collecting data: data sources, other sources

Briefly describe the following other sources of data an insurer can use for setting assumptions

Trade Magazines (3)

Actuarial Consultants (3)

Overseas Data (5)

Rate Table Software (3)

Answer 14

Trade Magazines

• national and global source, can perhaps serve as a reasonability check
• will lack level of detailed breakdown required

Actuarial Consultants

• national and international sources
• similar to reinsurer’s data, however consultants won’t share directly in insurer’s risk and reward.
• will usually charge a fee for service, and can be very expensive

Overseas Data

• may be of credible volume….
• …but different
  
  • culture, State healthcare, market practice, legislation and policy conditions
• may also be unreliable

Rate Table Software

• compares market premium rates for similar products of a given class.
• useful for gathering initial data or checking reasonability and competitiveness
• again, may not be able to obtain data at the correct level of detail

Question 15

Grouping the data

Once we’ve collected the data necessary for our investigation/to set assumptions, briefly list how grouping is done (2)

Answer 15

• divide data into relevant homogenous groups
• subject to adequate levels of data in each cell

Question 16

Adjusting the data

Give an overview of adjustments made to the data (3)

Answer 16

• data may need to be adjusted to meet the requirement of exercise being performed.
• more accurate allowances for adjustments can be made where insurer’s own experience is greater/grows with time
• unfortunately, the precision isn’t always obtainable when adjusting data => will need to make use of broad-based ratios to adjust data

Question 17

Adjusting the data: own data

Describe how data can be adjusted for the following (due to credibility and relevance)

Data from own experience (3)

Answer 17

Data from Own Experience:

• look to trends arising in data
• risk cells lacking credibility would be similarly weighted by market data
• can use standard tables or other methods

Question 18

Adjusting the data: local population

Describe how data can be adjusted for the following (due to credibility and relevance)

Data from local population (3)

Answer 18

Data from Local Population:

• insured lives different to population lives as insured are underwritten and less prone to claim.
• may also lead to moral hazard or anti-selection
• hard to say how they’ll differ

Question 19

Adjusting the data: reinsurer data

Describe how data can be adjusted for the following (due to credibility and relevance)

Data from reinsurers (3)

Answer 19

Data from Reinsurers:

• discuss relevance of data with reinsurer
• reinsurer may have already made allowances
• assess whether further loading is required, particularly for underwriting and claims

Question 20

Adjusting the data: market data

Describe how data can be adjusted for the following (due to credibility and relevance)

Data from market (4)

Answer 20

Data from Market:

• consider how own experience differs from market average.
• strictness of underwriting and claims might be reflected in similar ratios used for Standard Tables
• caution for launch of new product, strict underwriting may not mean smaller margins
• difference in policy conditions from average may produce straightforward ratios if analysed

Question 21

Adjusting the data: in depth

Describe how data can be adjusted for the following (due to credibility and relevance)

Data from overseas (2)

Answer 21

Data from Overseas:

• broad based population ratios to assess difference.
• many potential differences so requires a cautious approach.

Question 22

Claims incidence rates and recovery rates: calculating and adjusting rates, overview

(10)

(Basically an essay type question)

Answer 22

• Need to calculate rates for all product lines.
• Need suitable database, otherwise can calculate using exposed to risk techniques.
• Then need to adjust to produce relevant rates for time frame (likely use “own data” if available)
• Group data/rates into discrete/homogenous risk cells. Can then combine after analysis (consider credibility and regulatory restrictions)
• Where probability of claim is deemed to be a compound of many events, incidence rates are a combination of these probabilities e.g. PMI and CI. LTC usually single rate
• Allowance can be made for trends so that premiums reflect likely future outgo. Trends relevant if derived from market under consideration.
  
  • State healthcare plans/benefits are particularly important,
  • economic well-being/state of the country
  • inflation (high inflation will influence renewal premiums)
  • medical advancements over time should be allowed for
  • changing attitudes to health

Question 23

Claims incidence rates and recovery rates: calculating and adjusting rates, methods + adjustments

Comment on how rates may be calculated (2)

Comment on adjustments which can be made to rates (why adjustments are needed, and what these adjustments may be based on) (3)

Answer 23

Calculating rates:

• begin by choosing most suitable database ie reinsurer data, insurance industry data (eg mortaliy table).
• else, need to calculate rates using exposed to risk techniques ie for each homogenous risk group, sum of ‘decrements’ over exposure measure

Adjusting rates

• adjustments are made if a class of lives are expected to have experience different to that which observed from the data
• this could occur due to a change in the target market or distribution channel

Question 24

Claims incidence rates and recovery rates: calculating and adjusting rates, methods, datasource

Describe how to derive best estimate “base rates” a database is used (7)

Examples of databases include reinsurer data, or insurance industry data (such as a mortality table)

Answer 24

Base rates may be derived from adjustments done on data from a suitable database

• saves resources
• protects agains errors eg inappropriate graduation
• sufficient data: analysis own experience=> derive adjustmnts
• data must be over _enough years (_adequate volume), but also few enough years (prevent excessive heterogeneity from trends over time)
• analysis divide data into relevant homogenous groups
• further adjustments if different experience expected from that which analysed data relate (targ market, underwrit, distr)
• adjustments based on various data sources covered earlier in this chapter

Question 25

Claims incidence rates and recovery rates: calculating and adjusting rates, adjustments Describe how to derive best estimate morbidity and mortality rates, in terms of rates reflecting future experience (3)

Answer 25

1. Rates should reflect expected future experience of lives to be insured by contract being priced, in terms of * target market: affected by distribution channel * underwriting controls: * expected experience change: since last historical investigation, to point assumptions will apply on average (usually 10 - 15 yrs)

Question 26

Claims incidence rates and recovery rates: calculating and adjusting rates, adjustments, socio-economic factors Briefly desribe socio-eonomic factors which may influence the assumptions set for demographic rates used by insurers (in terms of the influence on the propensity of customers to consume healthcare) (4)

Answer 26

* Demand Increase: Changing attitudes, greater access, better diagnosis, new medical technology * Economic Pessimism and Unemployment: may encourage using PMI while employed before it is lost. Increases PMI incidence rates * High Inflation: impacts renewal price of PMI and may increase withdrawal rates, or policyholders choosing lower benefit levels. Also risk of selective lapses. * Government Welfare Provision: more generous State benefits, more difficult to retain business

Question 27

Claims incidence rates, variation: calculating and adjusting rates According to what may rates differ for CI, PMI and LTCI? CI (4) PMI (2) LTCI (1)

Answer 27

CI * CI rates are usually derived separately for each main class of claim eg heart attack, cancer, stroke, TPD, etc * Others are derived individually or proportionately. * Each projected separately and combined into global rate * Combining these separate rates is not appropriate for tiered benefits PMI * different claim incidence rates for different treatments for PMI. * sophisticated PMI models may allow for higher incidence rates among lives that have claimed historically and recovered. i.e. previous experienced used a risk factor LTC: * single rates will be derived for LTCI

Question 28

Claims incidence rates, risk factors: calculating and adjusting rates What risk factors are typically used for grouping of data/rates when setting assumptions for healthcare insurance contracts? (14) What restrictions might apply when using various factors to group data/rates? (2)

Answer 28

* Risk Factors which can be used to split data/rates: * age, * gender, * smoker status, * chronic condition, * occupation, * benefit size, * benefit type, * class of product (standalone/rider), * individual/group, * distribution source, * region * Also: * duration since entry, * duration since start of benefit, * underwriting status * In some countries, differentiation by one of more factors may be prohibited; but insurer can still monitor using such factors

Question 29

Claims recovery rates: calculating and adjusting rates For what kind of health insurance contracts are claim recovery rates particularly important? (1) How are claim recovery rates treated in practice when being estimated? (1)

Answer 29

* important where benefits are annuities and subject to continued incapacity e.g. LTCI (ceases on death or recovery). * recovery is usually unlikely once an individual needs LTC, so these rates are usually ignored

Question 30

Mortality rates: calculating and adjusting rates Under what circumstances are mortality rates quite important when modelling health and care contracts? (7) What data source should be used for mortality rates, and what is the general process of derivinv mortality rates comparable to? (2)

Answer 30

Mortality rates are important for: * for accelerated CI insurance * within the survival period for stand-alone CI insurance * among healthy lives for LTCI and CI * among lives in claim in LTCI * for claims in payment, the mortality assumption is less important pre-claim, unless a significant death benefit is provided * where policy does not have significant death benefits it is important that mortality is not overestimated as contract would be under-costed Data source + comparable process * Mortality rates should be taken from recent experience of credible body of policyholder data for the same contract * The process of deriving mortality assumptions is similar to that of deriving claim incidence and recovery rates.

Question 31

Persistency Outline how to set the persistency assumption when pricing a product (8)

Answer 31

1. Pricing method being used plays a role: persistency assumptions needed if using cashflow method, not if using formula method 2. Each healthcare contract wil exhibit different lapse expereince 1. PMI renewal will reflect rate of premium increase, service delivery, State healthcare plans and economic confidence 2. important for LTC contracts which offer surrender benefit 3. Lapses must reflect expected future experience of contracts to be taken out (eg full withdrawal, partial, and paid-up) 4. Based on analysis of company's recent experience, ideally of same contract, else of any similar contracts 5. If company doesn't itself have adequate data, there may be indutry-wide experience it could use 6. Assess results to see if they have been affected by special factors, e.g. adverse economic situation 7. Adjust for differences in class of lives, e.g. benefit changes, distribution channel changes 8. Sensitivity test/add appropriate margins * withdrawal rates significantly impacted by economic state/commercial factors which are diffiicult to predict. * therefore important to sensitivity test/add margins

Question 32

Persistency/lapse rates: influences and impacts What various factors can influence lapse rates experience by healthcare contracts? (4) Briefly describe some of the influences that lapses can have on healthcare products (4)

Answer 32

Lapses influenced by: * economic, political, commercial and awareness factors Lapses: * can cause loss where insufficient premiums received to cover initial strain (asset share negative) * can lead to profits: where no surrender value has been paid and reserve is accumulated, lapses are profitable * may be selective, can worsen risk pool as healthier lives leave. For PMI, lapse rates usually correlated with experience (lowest claiming members lapse more) * lapse and re-entry may be a problem, especially where competition drives down premiums

Question 33

Projection of Midpoint of Insurance Usage (5)

Answer 33

* Actuary will need to gather all info from local market on trends and developments to project. Consider future term for which current terms will be offered (shelf life of premium) * Incidence rates will vary by age and calendar year. * Claim termination rates will vary by duration of commencement from claim * Average claim size for PMI will be a function of inflation, current protocols and hospital charging structures. Will only need to look at a year at a time. * Where considerable uncertainty, allow for a review of premiums. Otherwise cost of reserves will make premium unaffordable.