Section C Flashcards
Life and Annuity Assumptions
MILES APP (iPhone app that talks to your running shoes)
- Mortality
- Interest rates
- Lapse rates
- Expenses
- Sales distribution and volume
- Average Size
- Premium persistency and payment pattern
- Policyholder behavior (option election, annuitization, etc.)
Experience Assumptions for Ind Life and Ann (LP-107-07)
Various Other Texts
Steps to Establish Experience Assumptions
- Identify the assumption required
- Determine the structure and each assumption
- Analyze experience and trends
- Review and adjust assumptions and reasonableness, consistency, and appropriateness
- Document assumptions
- Monitor experience and updated assumptions
Experience Assumptions for Ind Life and Ann (LP-107-07)
Expected Mortality: Fully Underwritten
Key principles in deciding the complexity of the assumption structure
- Differences in experiences assumptions should reflect differences in experience
- The definition of a class should be easily understood
- The number of classes should be practical and cost effective
Experience Assumptions for Ind Life and Ann (LP-107-07)
Experience Classes
When calculating an assumption, experience should be divided into experience classes
Experience classes will consist of contracts that:
- Are of similar type
- Have similar structure
- Are issued over a continuous period of time
- Have similar marketing objectives
Experience Assumptions for Ind Life and Ann (LP-107-07)
Analyzing Experience and Trends
USEERRs
- Use actual or similar experience
- Sensitivity test the assumptions
- Evaluate credibility of data
- Evaluate quality of data
- Reflect trends in experience as appropriate
- Reflect internal and external factors
Experience Assumptions for Ind Life and Ann (LP-107-07)
Assumption Documentation
- What the assumption is
a. Numerical values
b. Experience classes - Data underlying the assumption
- How the assumption was developed
a. Credibility method used
b. Any changes from prior study - How to use the assumption
Experience Assumptions for Ind Life and Ann (LP-107-07)
Structure of Mortality Assumption
- Type of mortality table: Aggregate, fully select, or select and ultimate
- ANB vs. ALB
a. LaTeX needed
b. LaTeX needed - Common experience
a. Age/duration
b. Gender
c. Tobacco use
d. Underwriting class (such as preferred, super preferred)
e. Policy size - Mortality Improvement
Experience Assumptions for Ind Life and Ann (LP-107-07)
Analyzing Mortality Experience
CRAM
- Credibility
- Risk covered
- Adjusting mortality for special situations
a. Multiple life policies
b. Substandard mortality
c. Term conversions
d. Anti-selection
e. Blending mortality tables
f. Adjusting similar experience - Mortality studies
Experience Assumptions for Ind Life and Ann (LP-107-07)
Credibility of Mortality Assumption
- Credibility is measured through a confidence interval (CI)
- 95% CI = m +/- 1.96 * s
a. m = n * q
b. s^2 = n * p * q - Enhancing credibility
a. Use multiple years of exposures
b. Group ages into 5 or 10 age groups
c. Could do an actual to expected analysis using an industry study - Expected value and variance for a group of policies – by counts
LaTeX needed
- Expected value and variance for a group of policies – by amounts
LaTeX needed
Experience Assumptions for Ind Life and Ann (LP-107-07)
Risks Covered in a Mortality Study
Only include standard risks in mortality study
Exclude the following:
- Policies not subject to normal underwriting
- Substandard policies
- Extended term or reduced paid up insurance
- Multiple life policies
Experience Assumptions for Ind Life and Ann (LP-107-07)
Mortality Studies
- Anniversary-to-anniversary or calendar year studies
- 5 years is the typical study period
- Amounts vs. Counts – use by amounts to reflect financial impact
- Mortality rate for a cell = Total Claims / Total Exposure
- Exact exposures
LaTeX needed
- Balducci assumption
LaTeX needed
- Common to make simplifying assumption that lives lapse at the end of the year and deaths occur in the middle of the year
Experience Assumptions for Ind Life and Ann (LP-107-07)
Term Conversion
- Term conversions usually result in higher mortality
- Ways to handle term conversions
a. Include in regular mortality studies – all permanent policyholders share in the extra cost
b. Include a charge in the term pricing - Formula for the cost of extra mortality s years after conversion
LaTex needed
Experience Assumptions for Ind Life and Ann (LP-107-07)
Conservations of Deaths
- Reflect anti-selection or new risk class in mortality experience
- The weighted average mortality of two populations balances back to the standard assumption
- One-year formula
LaTeX needed
- Multi-year formula
LaTeX needed
Experience Assumptions for Ind Life and Ann (LP-107-07)
Adjust Similar Experience to Reflect:
MR CUD (is chewing)
- Market - lower mortality for affluent markets
- Reinsurance quotes
- Company’s underwriting standards
- Underwriting classes and requirements
- Distribution channels
Experience Assumptions for Ind Life and Ann (LP-107-07)
Structure of Lapse Assumptions
- Common experience classes
a. Duration
b. Issue age
c. Frequency of premium
d. Policy size
e. Plan type (term vs. UL)
f. Marketing method
g. Target market - Shock lapse - may be applicable at end of surrender charge period
- Dynamic lapse - may be applicable for investment products
Experience Assumptions for Ind Life and Ann (LP-107-07)
Analyzing Lapse Experience
- Credibility is less of a concern for lapse rates when compared to mortality rates
- Lapse study – Almost identical to performing a mortality study
- Include these lapses in the lapse study:
a. Termination without value due to non-payment of premium
b. Cash surrenders
c. Transfers to extended term or reduced paid up - The lapse study may or may not include the following:
a. Term conversions
b. Partial withdrawals
c. Premium persistency
d. Termination because policy loan exceeds cash value
Experience Assumptions for Ind Life and Ann (LP-107-07)
Structure of Interest Rate Assumptions
- Deterministic, multiple deterministic scenarios, or stochastic scenarios
- Portfolio rates or New money rates
- Net Rate = Gross Rate - Spread
- Investment returns may be based on book value or market value
- Policy loans can be treated as an asset cash flow or a liability
Experience Assumptions for Ind Life and Ann (LP-107-07)
Analyzing Investment Experience
- Book value basis: I = 2I / (A + B - I)
- Market value basis: r = (B - A - C) / (A + C/2)
- Mutual fund returns: r = (B - A) / A
- Tim weighted return = LaTeX needed
- Dollar weighted return (R)
LaTeX needed
Experience Assumptions for Ind Life and Ann (LP-107-07)
Analyzing Expense Experience
- Expense assumption = Expenses / Units
- Units could be: Premium, policies, per 1000 of insurance, other
- Number of units in an expense study
a. If expenses are charged at the beginning of the year
Units = (A + B + N) / 2
b. If expenses are charged at the middle of the year:
Units = (A + B) / 2
- Expense allocation - indirect expenses need to be allocated down to the product level
Experience Assumptions for Ind Life and Ann (LP-107-07)
Projecting Expenses
- Historical expense trends
- Expected inflation rates
- Expected volume of business (economies of scale)
- Impact of any expected changes in the company’s business
- Productivity Gains
Experience Assumptions for Ind Life and Ann (LP-107-07)
Source of Data for Mortality Assumption
- Company Experience
- Inter-company experience
- Government or private sector population studies
- Medical studies
- Medical studies
- Private organizations, reinsurers, or actuarial organizations
Expected Mortality: Fully Underwritten
Criteria for a Good Credibility Method
- The method is practical to apply
- There is no double counting or omission
- All of the relevant information is used
- Results are reasonable in extreme cases
- The sub-category A/E ratios are reasonable when compared to company and industry experience
Expected Mortality: Fully Underwritten
Limited Fluctuation Credibility (LFCT)
- LFCT provides a method for establishing full and partial credibility
- Formula and notation
LaTeX needed - Formula for Z
LaTex needed
Expected Mortality: Fully Underwritten
Steps for the LFCT Normalized Method
- Calculate mortality ratios and credibility factors for the entire company and for each sub-category
- Calculate total company blended experience mortality ratio and expected claims using the credibility factor
- Calculate the sub-category blended experience mortality ratio and expected claims using sub-category credibility factors
- Normalize by multiplying each sub-category’s results by the ratio of the total expected claims in step 2 to the sum of expected claims in step 3.
Expected Mortality: Fully Underwritten
Advantages of the Normalized Method
- The sum of expected claims within sub-categories equals the total expected claims
- All of the information is used
- The results are reasonable in extreme cases
- The sub-category A/E ratios fall within the original range
- Correlations between subcategories may be captured
- It is practical to apply
Expected Mortality: Fully Underwritten
Adjust Mortality Experience for a New Underwriting Technique
Qnew = LaTeX needed QNew = the mortality rate adjusted for the new underwriting guideline Qold = the mortality rate based on past experience
A = Impairment frequency, the rate at which an underwriting technique will detect medical impairments
B = Sentinel frequency, the rate in which people with impairments will avoid the company due to the underwriting change
C = Additional mortality, the average amount that mortality is increased by a presence of the impairment
Expected Mortality: Fully Underwritten
Adjust Mortality Experience for a New Preferred Risk Class
- Qpref = Qstd * (1-B)
- QResidual = QStd * (1-A+B*A) / (1-A)
- A = The amount of applicants that will qualify for the preferred risk class
- B = The mortality differential for the preferred risk class
When adding multiple preferred risk classes start with the most restrictive risk class
Expected Mortality: Fully Underwritten
Considerations when determining selective lapse rate assumption
- Size of premium rate increase
- Period between premium increases
- Duration of policy
- Policy size
- Distribution system
- Heaped renewal commissions
- External market
- Proportion of healthy lives remaining
- Conversion activity
Expected Mortality: Fully Underwritten
Multiple Life Policies – Equivalent Single Age
- Mortality is approximated by using the mortality of a single age
- Rough approximation since joint life and single life mortality have different slopes
- Mortality rates for last survivor will be overstated in the initial durations
- Mortality rates for first to die will be understated in the initial durations
Expected Mortality: Fully Underwritten
Multiple Life Policies – Joint Equal Age
- Joint mortality is approximated by using the joint mortality of the same number of lives with the same age and underwriting class
- The JEA approach is superior to ESA
Expected Mortality: Fully Underwritten
Reasons why a flat percentage of an industry table may not be an accurate mortality assumption
MINimum mortality will be profitable
- Mortality improvement that varies by age
- Issue ages over 70
- New underwriting classes
Mortality Table Slope – The Discussion Goes On
Methods Used to End Mortality Tables
- Forced method
a. Set ultimate age mortality rate to 1.00
b. This creates a discontinuity - Blended method
a. Set ultimate age mortality rate to 1.00
b. Select an age less than the ultimate age and blend mortality rates gradually to 1.00 at the ultimate age - Pattern method
a. Let the pattern of mortality continue until the rate hits 1.00
b. The age at which the rate hits 1.00 is the ultimate age - Less-than-one-method
a. Select an ultimate age and DO NOT set the ultimate mortality rate to 1.00
Ending the Mortality Table
Challenges When Developing Term Mortality Assumptions
No Nobody LAPSE
- No ultimate experience past 10 years for new risk classes
- Nobody really knows how preferred/residual ratios change over time
- Lack of credible data
- Anti-selection after the level premium period
- Past mortality improvement may or may not continue
- Slope of mortality over the select period is unknown
- Each company has different underwriting criteria
Term Mortality and Lapses
Approached to Setting Mortality After the Level Premium Period
- Best Guess
- Becker-Kitsos (BK)
LaTeX needed - Dukes-Macdonald (DM) – n% is the amount of excess lapses that select against the insurance company
Term Mortality and Lapses
Lapse Experience Under Lapse Supported Products
- Ultimate lapse rates for lapse supported products are between 0.5% and 2%
- Term-to-100 products have very low ultimate lapse rates (<1%)
- Ultimate lapse rate for level COI UL policies is between 1-2%
Lapse Experience Under Lapse Supported Policies
Lapse Experience Under UL Level Cost