CH 13 (WM)

Question 1

State the general data considerations. [1.75]

Answer 1

Data should be:
- credible ✓
- relevant ✓
- readily available ✓
- reliable for the purpose intended.✓

The cost of obtaining/analysing the data and the format of the data should also be considered.✓✓✓

(CRrR)

Question 2

What does “relevance” mean in the context of data considerations? [0.75]

Answer 2

Relevance implies a similar class of lives✓ and policy conditions✓ to those expected for the product in question.✓

Additional marks: Own data is the most relevant source of data.

Question 3

What does “credibility” mean in the context of data considerations? [0.75]

Answer 3

Credibility implies a sufficiently large volume of data✓ to enable precise estimates to be calculated✓ (i.e. with relatively little sampling error).✓

Question 4

Describe why a health insurer would carry out data adjustments for credibility and relevance purposes? [1]

Answer 4

• The insurer will adjust data to meet the requirement of exercise being performed.✓✓
• More accurate allowances can be made where insurer’s own experience is greater.✓✓

Question 5

Describe how to make adjustments for projection to midpoint of insurance usage. [4.5]

Answer 5

• The initial estimates for assumptions are derived from past/historical data.✓✓
• The actuary will need to gather all information available from the local market on trends and developments ✓✓…
• in order to project this experience forward for the product’s lifetime✓.
• The actuary must incorporate not just the policy term, but also the length of time for which the current terms are to be offered.✓✓
• The assumptions will stay valid for an average of:
  [Expected shelf-life of proposed prem rates] +
  [Expected duration of policy to termination or to next review date].✓✓✓✓
• The adjustment method will be different for different assumptions.✓✓ Here we consider claim incidence rates, claim termination rates and average claim size.✓✓
• For example claim incidence rates vary by age and by calendar year✓✓, and
• claim termination rates vary by duration from commencement of claim.✓✓

Question 6

List the key assumptions to be analysed and how it will be projected for each of the main products. [2.75]

Answer 6

• LTCI✓ - claim inception rates✓
• CI✓- claim incidence rates by each claim cause then combined.✓✓
• PMI✓ - claim incidence rates and average claim costs for each benefit or treatment category separately.✓✓✓
  Rates should then be projected to midpoint of insurance usage.✓✓

(The question mentions “assumptions being projected” hence say something about projection of assumptions.)

Question 7

Claim incidence rates will vary by which factors? [0.5]

Answer 7

Incidence rates will vary by age and by calendar year. ✓✓

Question 8

Describe the adjustments required when projecting claim termination rates. [1.5]

Answer 8

Claim termination rates will vary by duration from commencement of claim.✓✓
Experience improvement or deterioration over time✓✓ may be assumed in producing expected claim rates, and claim termination rates.✓✓

Question 9

Describe the adjustments required when projecting the average claim size. [3]

Answer 9

For PMI the average claim size will be a function of:✓
* inflation✓
* current protocols✓
* hospital charging structure✓,
* as well as the availability of treatments and care under any state healthcare system;✓✓ but we need only look at a year at a time, because premium rates are typically reviewed annually, and policies are renewed once a year.✓✓
It is important that where there is considerable uncertainty as to future experience, the contract allows for some form of future review.✓✓
Otherwise the cost of the reserves could make the premium to be charged too high to sell, in terms of perceived value for money.✓✓

Tip: The “avg pol size” assumption is specific to PMI policies.

Question 10

State the “relevance factors” that should be considered when using external data for assumption setting? [3]

Answer 10

• Approach to UW and acceptance.✓✓
• Policy conditions✓, eg different diseases allowed as a CI claim by different companies.✓✓
• Claims management✓, eg different degree of effort made to rehabilitate LTCI claimants to get them back to care for themselves.✓✓
  -Distribution method and channels✓✓ leading to different types of lives insured.✓
• Target market✓

“example, different…” (x2),
eg different diseases allowed as a CI claim…&
eg different degree of effort…

Question 11

Describe the data “credibility issues” faced by an insurer and suggest how these shortcomings might be overcome. [2]

Answer 11

• For most HC business, the volume of business for many insurers over the recent past is not sufficient to be deemed credible.✓✓
• This is esp. true when the projection requires detail down to the level of particular risk cells.✓✓
• The actuary must then look elsewhere✓
• and must make appropriate adjustments✓✓
  to allow for the characteristics of external statistics, which may not be fully relevant as estimators of future claims outgo.✓✓

Question 12

Describe the suitability of “own data” for setting assumptions. [3]

Answer 12

• Most relevant source of data ✓✓
• but may not be credible, especially if data is broken into many risk cells. ✓✓
• may therefore need to be supplemented with external data. ✓✓
• Actuary may need to make adjustments to external data to make it appropriate. ✓✓
• Cost, reliability and format should not cause problems. ✓✓
• For new insurer the actuary will rely on other sources of data and not on internal as the latter will not exist. ✓✓

Question 13

Explain what is meant by population data. [1.75]

Answer 13

• Some national governments produce a periodic analysis of the healthcare experience for the population.✓✓
• The degree of sophistication can vary enormously.✓
• It is a very important source of information to the actuary in healthcare premium assessments✓✓; for some lines and some risk cells, it is the only base from which a first estimate of future costs can be established.✓✓

Question 14

Describe the suitability of “population data” for setting assumptions. [3]

Answer 14

• This will not be fully relevant to the insured lives or insurance products, eg claim definition.✓✓✓
• It should be very credible.✓✓

Disadvantages of population data:
- Questionable accuracy and reliability✓✓, esp. in HC where definitions may be subjective and the scope for double counting is significant.✓✓
- Data may not always be available electronically or in the appropriate format.✓✓
- May be out-of-date.✓

Question 15

Provide an example of double-counting in population data. [1]

Answer 15

• CI policies cover HA and CABG as separate causes of claim.✓✓
• PHs who suffer a heart attack as well as undergoing bypass surgery would only incur one insurance claim, but would be represented twice in the population records.✓✓

Question 16

Describe the two conflicting aspects when considering the relevance of national experience to that of insured lives. [3.5]

Answer 16

• Insured lives should be healthier and less prone to claim, ✓✓ because they have been underwritten and the less healthy declined or loaded. ✓✓
• However, moral hazard may be higher for PHs,✓✓ e.g. increased utilisation, i.t.o. both frequency and cost of treatment, ✓✓ of medical services when people are insured.
• Anti-selection can also be an issue. ✓✓
  If the policy is deemed non-essential, then people choosing to purchase insurance are arguably doing so because they expect to gain from the contract. ✓✓

Actuaries may choose to offer a small discount on population statistics but the validity of this will depend on the insurer’s approach to sales, marketing, UW and claims. [1]

Insured lives may be subjected to max age; also mostly employed lives.
acronym for last sentence - MUSC

Question 17

Describe the suitability of “reinsurance data” for setting assumptions. [2.5]

Answer 17

• The reinsurer should make the data as relevant as possible, ✓✓ although further adjustments may be needed for company’s specific underwriting & claims management. ✓✓
• Reinsurance premiums will be based on premiums calculated by insurer hence determine treaty profits. ✓✓
• Credibility will vary, but more credible than own data. ✓✓
• Reinsurer will likely provide data in return for a share of business. ✓✓

Question 18

Describe the suitability of “insured lives data” for setting assumptions. [2.5]

Answer 18

• It is generally of sufficient volume to be statistically credible.✓✓
• This should be more relevant than population data as it is based on local experience of insured lives.✓✓
• It has been compiled by experts so accuracy should not be an issue.✓✓
• Drawback is that it is a market average; doesn’t relate to any one insurer.✓✓
• May not be relevant to individual insurers and by time it is published, may be out of date.✓✓

Question 19

For insured lives data explain an approach actuary can follow to make adjustments for the market average and company’s own experience. [1.5]

Answer 19

• The strictness of UW and claims management might be reflected in similar ratios used for life insurance standard tables. ✓✓
• But caution should be taken if this is a new launch into a new HC product; ✓✓
• strict UW for mortality may not translate directly into similar margins when UW for morbidity for CI business. ✓✓

Question 20

Describe the suitability of data obtained from returns to the insurance supervisor for setting assumptions. [3.75]

Answer 20

• These are likely to be highly accurate. ✓
• The degree of detail that the actuary can extract for the estimation of future risk in individual cells is questionable though. ✓✓
• The returns are produced to assess solvency and certainly not to assist competitors with pricing. ✓✓
• Can provide a high-level check on premiums produced from other sources. ✓✓
• The data has national relevance to company concerned. ✓✓
• However care should be exercised since policy conditions might differ. ✓✓
• Approaches to UW & acceptance and claims management procedures are likely to be materially different from the market average. ✓✓
• This data should be more credible than insurer’s own data but less credible than population data. ✓✓

Question 21

List all the sources of data [2.25]

Answer 21

• Own data
• Population Data
• Reinsurers data
• Insured lives data
• Returns to the insurance supervisor
• Overseas data
• Actuarial consultants
• Rate table software
• Trade magazines

Question 22

An insurer has some overseas insured lives data for a particular product, but no home industry data.
Provide a formula that shows how rates for the home insured lives can be derived. [1]

Answer 22

{home insured industry morbidity rates}/{o/s insured industry morbidity rates} =

{home population morbidity rates}/{o/s population morbidity rates}

Question 23

List the adjustments for relevance purposes that are required when using overseas data. [1]

Answer 23

Adjustments required for differences in:
- State HC provision
- Market practices
- Legislation
- Policy conditions

(Acronym: SMaL P)

Question 24

Describe the suitability of “data from actuarial consultants” for setting assumptions. [2.5]

Answer 24

• The consultant/actuary should make the data as relevant as possible, ✓✓ although further adjustments may be needed for company’s specific underwriting & claims management. ✓✓
• Credibility will vary (perhaps check this with consultant?), but more credible than own data. ✓✓
• Consultants would normally charge a fee for their service. ✓✓
• However likely to have less of an interest in risk. ✓✓

Question 25

Discuss the merits of data from rate table software. [1.5]

Answer 25

• Useful in gathering initial data, ✓✓
• or checking the reasonableness and competitiveness of final premiums rates. ✓✓
• Very similar to data from trade magazines, but usually in more detail. ✓✓

-

Question 26

Discuss the merits of trade magazine data. [2]

Answer 26

Lacks detailed breakdown, but they may provide market information✓✓

and a reasonability check.✓✓

Will give information on the level of benefits offered and premium rates charged by other companies.✓✓

They will not, however, normally include any claim statistics.✓✓

(“trade” data so must say something about market info - 1st bullet)

Question 27

List the data requirements to calculate claim incidence rates for standalone CI policies. [4.75]

Answer 27

Exposure data: ✓
* dates of birth, entry, exit, diagnosis or treatment, notification of sickness, claim payment.
* reason for exit, type of disease or procedure involved. [2]

This allows calculation of central exposures to risk, grouped by age, for healthy PHs; Ex(H). [1]

Claims data: ✓
- The number of claims due to each particular disease or procedure arising from the exposure✓✓,
also grouped by age, Ix (j) [relating to the j’th disease/procedure type] should be counted.✓✓✓

Question 28

List the data requirements to calculate claim inception rates for LTCI policies. [4.75]

Answer 28

Exposure data: ✓
* Dates of birth, exit, diagnosis, first becoming incapacitated, notification of incapacity, start of LTC claim, end of LTC claim. [1.75]
* Reasons for exit and/or end of LTC claim. ✓✓

This allows calculation of central exposures to risk, grouped by age, separately for healthy and for claiming PHs, separately for each of the different claim states; Ex(A) and Ex(L). ✓✓

Claims data: ✓
- Count the number of transitions arising from exposure, also grouped by age. ✓✓
- This will additionally require dates of transition from one benefit level to another, but no return transitions will normally occur. ✓✓
- Make appropriate allowances for deferred periods. ✓✓

Question 29

List the data requirements to calculate claim incidence rates and average cost of each claim for PMI policies. [1.75]

Answer 29

Assuming all policies are annually renewable, ✓ we need to:
* count all policies that are taken out or renewed during each year of the investigation, subdivided by age; Nx. ✓✓✓
* count the number of claim incidences arising from each year’s cohort of policies, subdivided by age and class of benefit or procedure involved; Ix(k), where k is the benefit class. ✓✓✓
* record the total amount of all claims paid from each cohort; Cx(k). ✓✓

**(We dont need to calculate exact exposures because of the 1 year policy term.)

Question 30

List the two sources of market data. [0.5]

Answer 30

• Insured lives data
• Published Returns

Question 31

Explain what is meant by reinsurance data. [1.5]

Answer 31

• A reinsurer often provides statistics relevant to the risk to be insured✓✓ (to assist in pricing and reserving)✓ as part of the range of services on offer.✓
• The reinsurer will be keen that the statistics proposed are as accurate as possible, since the profits under the treaty will reflect the premiums calculated.✓✓

Tip: The word statistics appear in each of the two sentences.”

Question 32

Explain what is meant by “Insured Lives” data. [2]

Answer 32

• In some well-established markets, there may be an agreement between companies to pool claims and policy statistics,✓✓
• so that overall industry-wide information might be available that would be more credible than that which individual companies could access.✓✓
• An example of this is the CMI in the UK which collects data from nearly all insurers in the CI insurance markets.✓✓
• The compiled agregate information is published for the benefit of all.✓✓

Tip: Standard Tables

Question 33

Explain what is meant by “returns to an insurance supervisor” data. [2.75]

Answer 33

• In most territories, insurers have to submit returns to the local regulator at regular intervals.✓✓
• These returns show much that is common to normal company accounts✓, incl. information on annual revenue and BS items.✓✓
• In addition they will typically incl. much more detailed info about the policies in force:✓✓ primarily this has to do with the calculation of the supervisory reserves✓, but also include info on the numbers of policies, amounts of premiums and benefits, all broken down by policy class.✓✓✓

Question 34

Outline how you would derive claim incidence rates for CI policies from the Exposure Data, Ex(H) and Claims Data, lx(j). [5]

Answer 34

All policies showing dates of diagnosis occurring in the investigation period✓✓, and which ultimately lead to a claim payment✓, should be included✓. This should be the case even if the date of claim payment occurs after the investigation period.✓✓

The Ix(j) figure should include an allowance for any incurred but not yet settled (IBNS) claims.✓✓

IBNS includes those claims occurring during the investigation period that, at the time of performing our investigation, have been reported to the company but have not yet been paid out✓✓ (these are sometimes called reported but not yet settled – RBNS)✓. It also includes future claims arising from diseases commencing during the investigation period but which have not yet been reported (called incurred but not reported – IBNR).✓✓

The Ix(j) figure should exclude all claim payments made during the investigation period that relate to diseases diagnosed (or procedures carried out) prior to the start of the period.✓✓

The claim inception transition rate for disease / procedure j is then estimated from: M j (x+1/2) = Ix(j) / Ex(H) ✓✓✓✓

Initial (q-type) rates of transition could then be calculated, if required, using (for example)✓:

qx(j) = 1-exp-[M j (x+1/2)] ✓✓

Question 35

Describe the advantages and disadvantages of using market data as opposed to that derived from the in-house portfolio, for the purpose of analysing trends. [4]

Answer 35

Advantages
It will be drawn from a wider pool of experience and so trends may be spotted earlier.✓✓ Trends in your own data might be put down to random fluctuation.✓
Your product may be new and you might not have any experience at all.✓✓

Disadvantages
The experience of others may not be applicable to your own experience, for e.g. due to different policy designs and u/w conditions.✓✓✓
There will be delays between the POI and the date of publication of market data.✓✓
Market data may be heterogeneous and observed trends may therefore be spurious.✓✓
Not all the market participants will be included.✓
Suitable data may not be available in the right format or at the right time.✓✓
Market data may be subject to errors and insurer will not be able to verify the data.✓✓

Question 36

(a) List the sources of data the actuary may typically use in claims experience investigations.
(b) Describe the conflicting factors that actuary need to consider when dividing data to credible cell levels, together with how the various data sources will be used for analysing the experience at a cell level. [4]

Answer 36

a) Market data/Industry data✓; Reinsurers✓ ; Competitors✓ ; Similar contracts✓

b) The actuary will have to balance the need for subdivision of information into risk cohorts that are subject to homogenous risk experience, against the need for adequate data to provide meaningful results.✓✓

Own experience will be analysed initially at a portfolio level, prior to greater subdivision.✓✓
The actuary will look to market/population data or reinsurer information on an ongoing basis to provide insights at a cohort level✓✓ and analyse the comparative risk levels.✓

Question 37

What are the advantages and disadvantages of using reinsurer’s information? [4]

Answer 37

Advantages
* Reinsurers have access to health experience of many direct writers.✓✓
* Sometimes the reinsurance data may be the most relevant available.✓
* This is most likely to be the case for a new insurer that will have little data of its own, or in the case of new or substantially revised contract. [1]

Disadvantages
* It relates to a larger number of companies.✓ Each of these companies will have their own target market, and u/w and claims procedures.✓✓
* There may be confidentiality issues in the amount of detail that the reinsurer can provide, esp. if the reinsurer is closely connected with one major insurer in the market.✓✓
* In some circumstances, even the reinsurer may have little or no suitable data.✓✓
* The reinsurer will almost certainly want something other than gratitude for providing assistance. But if aim is to reinsure anyway, then there may be no real loss.✓✓✓
* However, if reinsurer’s technical expertise forces you to reinsure more than you would otherwise choose to do, the cost of that extra reinsurance is a problem.✓✓

Question 38

Describe the use of the various sources of information when estimating future claim trends. [1.75]

Answer 38

The actuary will use various sources of data to improve the estimation of future experience.✓✓ The actuary will use:
* market intelligence✓ and
* published population data✓
* in addition to that derived from own (in-house) data✓

…in order to discern trends in
buying attitudes✓
and more particularly in the propensity to claim.✓

Question 39

Describe the factors that actuary need to consider when comparing results from claim experience to those of competitors. [2.5]

Answer 39

The actuary’s analysis of competitor products and processes is a necessary part of experience monitoring.✓✓
A competitor’s approach to risk selection in pricing can present problems when the competitor is subdividing the population by more rating factors✓✓and thus altering the likely population over which the first company’s business will be averaged;✓ this will amount to selection of the better risks and the first company’s experience will deteriorate.✓✓
It can also present opportunities where the market has failed to recognise key features of the risk✓✓ and thus a more profitable experience can be achieved by more appropriate segmentation and pricing.✓✓

Question 40

State the reasons why adjustments might be required to overseas data. [4.5]

Answer 40

• Different subsets of lives
  (relative to the population) take out insurance policies in the two countries✓✓, which may be a consequence of different:
  
  • selling methods✓
  • cultures✓
  • State healthcare provision✓
• different policy conditions will make claim rates different, even if the class of lives is equivalent✓✓
• the overseas population morbidity data may be compiled differently from home population statistics✓, using different disease criteria or groupings✓✓, making comparison inappropriate✓
• underwriting standards may be different✓
• claims management procedures may be different✓
• market practices, eg relating to claims disputes, may be different✓✓
• legislation / regulation✓, eg on rating and acceptance of policies, may be different✓✓
• the data may not relate to the same time periods✓✓
• the overseas industry data are likely to be very heterogeneous, so it is unlikely that any single company’s experience will reflect the industry average✓✓

Question 41

Suppose we have estimated claim incidence rates that apply, on average, on 1 January 2004 (so 2004 is our “base” year), for lives aged x.

We are planning to launch a product on 1 January 2005, for which we hope the premium rates will not need to be changed for three years.

Explain how you woluld estimate the premium rate for lives aged x at entry from the base rates. [4]

Answer 41

The midpoint of insurance usage is 1 July 2006.✓✓

So the first policy year is from 1 Jan 2006 until 31 Dec 2006.✓

The exposure for the 1st PY is the average for the lives aged x at the start (1 Jan 2006) and the end of the first policy year (1 Jan 2007), respectively.✓✓

The based rates for 2004 are defined as ix(0).✓✓

So for a life aged x at entry, we would calculate their premium rate assuming claim incidence rates of:

1/2[ix(2) + ix(3)] in the first PY ✓✓✓✓ ,and
1/2[ix+1(3) + ix+1(4)] in the second PY. ✓✓✓✓
… and so on.✓