Monitoring the Emerging Experience Flashcards
Why analyse experience and trends?
To understand drivers of experience
To identify corrective actions
Develop a history for modelling
To provide information to management
To review and update your model
What are the main things analysed?
Because of time and resource constraints: we analyse factors with a large financial impact or a large impact adversely. Look at Actual vs expected experience.
Can see what has large financial impact by sensitivity testing
Minor deviations can be analysed less frequently
What are causes of minor vs large deviations
Often ‘minor deviations’ are caused by demographic factors while major deviations arise from economic
Why can claims experience sometimes be tricky to analyse Actual vs Expected
The time frame of the review might be too short so there might only be
limited decrements. Lack of data
May have too few data points on claims, etc. to give a statistically credible graduation of the experience.
What are the main actuarial decrements?
mortality, morbidity, claim experience.
What are two main purposes of analyzing experience
Two purposes of analysing experience can be for management reporting on profit and for modelling improvements.
Management will often want analysis of profit by categories ex: distribution channel, product grouping, geographical area (which will mp to one manager)
Name the types of experience to analyse
Demographic, economic, expenses, business volume, profit and return on capital
Note there are links between these
Give examples of demographic experience and business volume experience things that could be analysed
Demographic: Claims experience, persistency, lapses, surrenders
Business Volume: Conversion rate of: How many potential customers turn into actual customers.
Give an example of links between different categories of experience
Business volume and many expenses - more business means more expenses but also should mean more profit
Profit and all of the other headings
What are some sources of deviations in actual vs expected
Mortality, morbidity, withdrawal or lapse - in number or amount
Investment income and gains
Expenses
Inflation
Salary growth
Taxation
Premiums or contributions paid
New business levels
What are the key sensitivities in business model for general and life insurers?
General insurance: Key item is claims experience
Life Insurance: Hard to know if its risk or investment business.
General insurance: Key item is claims experience
For risk element: Claims experience and expenses are key
For investment business it is expense and persistence that tend to be crucial.
What are the key sensitivities in the business model for pension funds
Relationship between salary escalation and investment performance over the period for DB Pension Funds has direct effect on liabilities.
What are the key sensitivities in the business model for banking and health insurance
Banking: default on loans and pricing of credit risk (borrowing vs lending spread)
Health insurance: persistence and claim rates, including (medical) inflation in claim costs.
What is significant about pricing of health insurance in ireland?
Insurers in Ireland have premiums set which can only be changed based on their age so everyone aged 30 has to get the same pricing. The insurers avoid this laws by having lots of products with slight tweaks.
How to analyze experience: Process
- Establish the objectives
- Collect the data
- Analysis of Actual to Expected
- Validate results
- Report the results:
How can validating results be done?
Validating results is key but can be difficult: Could use consultants, industry data, recent hires from other companies.
How can one get an idea for expected experience
Expected experience is needed to analyse experience
A way to do this is to use a cash flow model, modelling all decrements
We also need all this original analysis to detail all assumptions made.
How might one analyse the demographic assumptions
Calculate the expected outcome, under best estimate assumptions in the model. Then compare actual claims with expected or actual lapses with expected.
Compare Actual to expected rates. But generally cannot estimate reliable rates as data too sparse - may need to group data.
Be aware: unusual outcomes whether good or bad could be an issue
Actuarial models typically have demographic assumptions rated by age, gender, and duration and other
What issues can arise when analysing the demographic assumptions
Incurred claims/lapses etc that are not reported can occur depending on the timing of notifications.
So some ‘appropriate’ time after period ends is needed before investigation AND make an explicit IBNR estimate
Should also compare long run trends and compare to your demographic experience.
Question arises: How much sub division can data support Balance between credibility and homogeneity.
What issues can arise because of data
How much sub division can data support Balance between credibility and homogeneity.
Data needed is : financial amount of occurrences AND the exposed-to-risk for that incidence. Mya have to wait a long time to get this or settle for less accurate analysis.
Audited data is more reliable
If grouping of data is used issues can arise if there is change in the composition fo the exposed to risk (so cannot aggregate data):
Underwriting standards changing with time
Selective lapse experience changing with time
Average premium size changing with time
Policies from different sales channels