Reserving Text Flashcards
Flashcards copied from Reserving Text
Describe 5 components of unpaid claims
1) Case outstanding (on known claims): this reserve is set by the claims department, third party adjustors, or independent adjustors for known and reported claims only (the future expected payments for claims already reported.
2) IBNER: a provision for future development on known (case) reserves (incurred but not enough reported).
3) Reopen: a provision for expected costs due to the reopening of claims.
4) IBNR: provision for claims incurred but not reported (this is the narrow definition of IBNR). IBNR comes about due the lag between a covered event occurring, and it being reported to the company—in some lines of business this time lag can be very long.
5) In transit: a provision for claims reported, but not yet recorded within the database/system.
How do you calculate incremental paid claims, case outstanding, and reported claims?
Incremental paid claims during a period are the sum of payments in the period, or equivalently cumulative paid claims at the end of the period minus cumulative paid claims at the beginning of the period.
Similarly, incremental case O/S changes during a period are the sum of changes in case O/S in the period, or case O/S at the end of the period minus case O/S at the beginning of the period.
Incremental reported claims within a period can be calculated in the following ways:
1) Reported claims at end of period – reported claims at beginning of period.
2) Incremental paid + case O/S at the end of period - case O/S at the beginning of period.
3) Incremental paid + incremental case O/S.
4) Sum of incremental reported claims during the period.
External data commonly used when (3):
1) Small company: data sets not credible.
2) Systems limitations limit availability of data.
3) Entering a new line of business or region.
External data is particularly useful as a benchmark for (3):
External information can be very valuable as a benchmark, especially for the following:
1) Tail development factors.
2) Trend rates.
3) Expected claims ratios.
External data may be misleading or irrelevant due to differences in (5):
External data may be misleading or irrelevant due to differences in:
1) Insurance products.
2) Case outstanding and settlement practices.
3) Insurer’s operations or coding.
4) Geographic areas
5) Mix of business/products
Consideration for data groupings for an unpaid claims analysis (6):
1) Volume of claim counts in the group.
2) Length of time to report the claim once an insured event has occurred- i.e. reporting patterns.
3) Ability to develop an appropriate case outstanding estimate from earliest report through the life of the claim- i.e. accuracy of case reserves.
4) Time to settle the claim once it is reported- i.e. settlement, or payment patterns.
5) Likelihood of claims to reopen once it is settled.
6) Average settlement value- i.e. severity. -Must balance credibility and homogeneity of groupings. If relative volume of business between two or more segments is constant, less homogeneity in groupings is needed.
Treatments of ALAE under excess of loss reinsurance (3):
1) ALAE included with claims in the limit.
2) ALAE not covered.
3) ALAE covered pro rata: i.e. covered in the same proportion as claims, e.g. if $5M of a $15M claim is covered, 1/3 of ALAE will be covered.
Examples of exposures used in estimating unpaid claims: (8):
1) Earned premium: most commonly used, on-level if available.
2) Written premium.
3) Policies in force.
4) Number of vehicles insured- auto.
5) Payroll- WC.
6) Property value- property insurance.
7) Number of employees- crime.
8) Sales- general liability.
Define the following: Policy effective dates Accident date Report date Accounting Date Valuation Date
Policy effective dates: policy begin and end dates.
Accident date: the date of covered occurrence.
Report date: date claim reported to the insurer.
Accounting date: the date through which liabilities are estimated, for all claims occurring on or before regardless if they have been reported or not.
Valuation date: the date through which known claim data is included, regardless of when the actuary performs the analysis.
Describe Calendar Year aggregation (including Advantages and Disadvantages)
Calendar Year (CY): payments include all claims paid in a given year regardless of when the accident, policy, or report of claim occurred. CY exposures (e.g. calendar year earned premium) are commonly used, but CY claims are rarely used to estimate unpaid claims. CY earned premium can be calculated as: WP + (Beginning Unearned Premium Reserve - Ending Unearned Premium Reserve).
Advantage: readily available and is not subject to further development. -
Disadvantage: cannot be used it to estimate development of claims.
Describe Accident Year aggregation (including Advantages and Disadvantages)
Accident Year (AY): by far the most commonly used aggregation for claims data. Claims are grouped by the year of occurrence, regardless of when the claim is reported, policy was written, or payments are made. Actuaries commonly use CY exposure data with AY claims data (CY exposures approximately match AY claims).
Advantage: Represents a shorter timeframe than policy years (or underwriting years) so we can estimate ultimate claims sooner, and there are numerous industry benchmarks available on an AY basis.
Disadvantages: slight mismatch with CY exposures, the potential for mixing of claims from policies underwritten/priced at differing levels, or at different retentions for self-insureds.
Describe Policy Year aggregation (including Advantages and Disadvantages)
Policy Year (PY) or Underwriting Year: Claims can be grouped by when the policy was written (PY if a primary policy, U/W year if a reinsurance policy) regardless of when claims occur, are reported, or paid. -
Advantages: PY claims directly match PY exposures, and are useful If there have been significant changes in policy characteristics or shifts in types of business written from one year to the next (since AY aggregation will mix claims from different PYs).
Disadvantages: the extended timeframe. it can also be hard to understand or isolate the effects of a large event (e.g. catastrophe). If a catastrophic event occurs in April, it will affect policies written in the prior year and current year (i.e. the effect will be spread over the current and prior PY).
Describe Report Year aggregation (including Advantages and Disadvantages)
Report Year (RY): Claims can be aggregated by the year claims are reported, regardless of when the claim occurred, are paid, or when the policy was in effect.
Advantage: number of claims is fixed by the end of the first year, resulting in more stable data.
Disadvantage: only measures development on case (IBNER) rather than newly reported claims (IBNR).
Sample Topics/Question with: Claims Department
Claims: understand any changes in claim settlement or report speed, changes in case adequacy (or how case reserves are set), shifts in large/small claims, or changes in the rigor of claims defense.
Sample Topics/Question with: Underwriting Department
Underwriting: understand shifts the book of business, large risks underwritten, and pricing of policies.
Sample Topics/Question with: Data/Accounting Department
Data or Accounting: understand changes in claim coding, data availability and definitions, and processes in place to ensure data accuracy.
Sample Topics/Question with: Ratemaking Actuaries
Ratemaking actuaries: understand any changes in operations, shifts in the book of business, use and availability of external data, filing information, and priced claims ratios.
Sample Topics/Question with: Reserving Department (if you are a consultant)
If you are a consultant and meeting with in house reserving actuaries, ask for prior reserve reports to review, ask about areas of disagreement between prior studies, and for any additional background they can provide.
Sample Topics/Question with: Reinsurance Department Senior Management
Reinsurance: details of the reinsurance programs, especially any changes in the program or retentions.
Sample Topics/Question with: Senior Management
Senior management: company overview, types of business written, types of marketing and distribution channels (e.g. direct writer, independent agents, captive agents), and any organization changes.
Describe the 3 important dimensions of the development triangle
Rows, the columns, and the diagonals.
Each row in the triangle represents one accident year.
Each column represents the age (or maturity) of development.
Each diagonal represents the values at each successive valuation date.
The sum of the latest diagonal minus the sum of the second latest diagonal represents the amount paid during that given timeframe since each valuation represents a cumulative value.
Explain how for self-insurers the policy year, fiscal year, and accident year are commonly the same
Self-insurers issue policies on a single date (rather than throughout the year as insurers do) and this date often corresponds to the first day in the fiscal year, and in these cases the accident year is defined over the same timeframe as the fiscal year (accident years do not necessarily have to run 1/1 to 12/31). In such a situation the policy year, fiscal year, and accident year would all be the same timeframe.
What should be considered when deciding on the number of periods to include in the data triangle?
Line of business, state, and type of data. Ideally, we would like to have development periods available until we expect no further development– often we will not have this amount of data and will need to use tail factors to project from the oldest maturity to ultimate.
What are 4 very basic diagnostic checks that should be done for claims amounts?
1) Visual examination of the Reported claims development triangle.
2) Visual examination of the Paid claims development triangle
3) Triangle of Reported Claims/Earned Premium (on-level if available).
4) Triangle of Paid Claims/ Earned Premium (on-level if available).
What diagnostic triangle is useful in looking for changes in case outstanding adequacy or settlement patterns?
The actuary can examine the consistency of paid claims to reported claims by looking at a triangle of paid to reported claims.
If case reserves are becoming more adequate the ratio would go down because case is contained in reported claims in the denominator.
If settlement patterns are increasing, we would expect the ratio to increase at earlier ages since more will be paid earlier and paid claims are in the numerator.
Since this diagnostic is a ratio, a change in either could be causing the differences, and no change in the ratio does not necessarily imply there are no changes in either as there could be offsetting changes in each.
When evaluating a diagnostic triangles what dimension are we generally looking at?
We are generally most interested in changes down columns– differences by AY at the same maturity indicting changes in development patterns.
What are 3 very basic diagnostic checks that should be done for claim counts?
1) Visual examination of the Reported claim count development triangle.
2) Visual examination of the Paid claim count development triangle
3) Triangle of Closed Counts/Reported Counts—Examining the ratio of closed counts to reported counts is a very useful metric which can indicate a change in the settlement rate of claims. Blips in this ratio can be caused by one time internal or external events (e.g. systems outages or catastrophes)— we are more interested in persistent changes or trends.
What are 3 diagnostic checks use averages? What information can we gain by comparing them?
1) Triangle of Reported Claims/Reported Counts (i.e. avg. reported = reported severities).
2) Triangle of Paid Claims/Closed Counts (i.e. avg. paid = paid severities).
3) Triangle of [Reported– Paid Claims] / [Reported -Paid Counts] (i.e. avg. case outstanding).
We are interested in comparing the trends down the columns of each. If things are steady trends would be fairly similar.
If case adequacy is increasing we would see higher trends in the average outstanding and average reported triangles than the average paid claims triangle.
We can also look for shifts within each, indicating a change in type of claim being settled at each age.
What the basic steps to complete the development technique? Why will the latest diagonal of the age-age triangle be blank?
1) Compile claims data in a development triangle (as covered in Chapter 5).
2) Calculate age-to-age factors (aka link ratios or report-to-report factors).
3) Calculate averages of the age-to-age factors.
4) Select claim development factors (CDFs).
5) Select a tail factor.
6) Calculate cumulative claim development factors.
7) Project ultimate claims.
Due to the shape of a data triangle, the triangle of age-to-age factors will have one less row and column than the data triangle (i.e., the latest diagonal will not have an age-age factor since the following period has not yet occurred).
What characteristics should the actuary look for when selecting claim development factors?
1) Smooth downward progression of age-age factors across development periods.
2) Stable age-age factors within development periods: in general, less mature age-age factors are often less stable– there should be no clear trends down columns, even at less mature ages (i.e. early volatility should be random).
3) Credibility of the experience: if the insurers own historical experience data is limited the actuary may use experience from similar lines with similar claims handling practices within the company, or use industry development factors.
4) Changes in patterns: review age-age factors for changes that may suggest changes in internal operations or external environment such as trends in age-age factors down the columns.
5) Applicability of historic patterns to the future: consider changes in the book of business, insurer operations, and external factors both observed in the historic period and those that may not have shown up in the claims experience yet.
3 approaches to estimate tail factors
1) Reliance on industry benchmarks.
2) Fit a curve to the selected (or observed) development factors.
3) If reported claims are assumed to be at ultimate value, use the ratio of reported to paid claims at the oldest age as the tail factor of the paid triangle (since reported claims are assumed to be at ultimate value, the ratio measures paid development to ultimate value). The reported tail will be 1.00 since it is assumed to be at ultimate already.
% Reported % Paid
XXXXXXXXXXXXXXXXXXXX Rsv. Chapter 7 – Development Technique
General observations and common relationships of the development technique
Claim development factors will generally be larger for the most recent (most immature) accident years and smaller for older (more mature) accident years. As a result, IBNR will often be largest in the most recent accident years. This makes sense because as the accident years mature more claims will be reported and paid, moving the IBNR and unpaid loss estimate closer to 0.
Development factors tend to increase as retentions (limits in the underlying data) increase because large claims tend to be reported later, thus at a higher retentions (limits) larger losses will be observed later in the triangle resulting in higher claims development factors.
As such, retention must be taken into consideration in estimating ultimate losses and deciding how to segment your data into triangles. Ignoring retentions will lead to errors in both reserving and ratemaking.
What is primary assumption of the development technique? Implicit assumptions when using paid and reported claims? When it is appropriate to use the development method?
Method assumes we can predict future claims activity based on historical claims activity to date by looking at how prior periods developed– primary assumption is that reporting and payment patterns will be similar going forward to patterns in the past. Using reported claims implicitly assumes there have been no significant changes in case adequacy levels during the experience period. Using paid claims implicitly assumes no material changes in speed of claim closure and payment. Appropriate for insurers in a relatively stable environment with no major organizational (e.g., new claims processing system) or external environmental (e.g., tort reform) changes. Requires a large volume of historical data, and works best when data is not distorted by large claims. The method works best with high frequency low severity claims. ? If assumptions do not hold, the actuary should consider alterative techniques (or at least adjust the selected claim development factors).
Additional implicit assumptions of the development method Advantages of the development method Disadvantages of the development method
The method also implicitly assumes a stable mix of claim types, stable policy limits and reinsurance retentions, consistent claims processing, and that claims are evenly spread throughout the periods being evaluated. Advantages: it is the most basic method, it works well if the book of business is stable, especially for low severity high frequency lines of business. The CDFs and claim estimates are inputs to many other methods. Disadvantages: is very susceptible to changes in the book of business or environment, and is very sensitive to claims reported (paid) to date (i.e., estimates are often highly leveraged).
3 examples of ways an actuary can set the a-priori ultimate claim estimate
1) Multiply the ratemaking expected claims ratio (ECR) by earned premium in the period.
2) Estimate expected claims using a complex simulation model requiring extensive expert input (No specifics are needed for the exam, this is just given as an example).
3) Estimate the claims ratios for each AY by developing, trending, and adjusting actual claims from each other AY for benefit level changes, then dividing by on-level earned premium. This is the method used in most of the examples in the text and tested the most.
3 situations the expected claims method is commonly used
1) When entering a new line of business or territory, therefore data is unavailable for other methods. 2) Historical data becomes irrelevant for projecting future claims, due to changes. 3) Very long tailed lines of business when the development factors are highly leveraged (high CDFs). 4) As an input to the BF method. Rsv. Chapter 8 – Expected Claims Technique
Describe the steps to complete the textbook expected claim technique
1) Develop an initial ultimate claim estimate—commonly the average of the paid and reported development estimates.
2) Adjust premium and claims to the target year cost levels (on-level premiums, trend and adjust claims for tort reform, trend exposure base if using something other than premium that is inflation sensitive).
3) Divide the adjusted claims by adjusted premiums to get claims ratios (CR) at the target years’ level—make a CR selection (exclude the data of the target year itself from the average you use for selection).
4) The expected claims method ultimate loss estimate (a-priori ultimate claims) for that AY are then the selected CR times earned premium.
5) Repeat for other immature AYs— CRs for mature AYs will often be selected judgmentally or together.
Advantage of the expected claims method Disadvantage of the expected claims method
Advantage: stability
Disadvantage: also stability, or more specifically lack or responsiveness—the actuary can make it more responsive by actively monitoring and updating the a-priori estimation as needed
