Part 5 Practice Exam 2

SOLUTIONS
1. (1.5 points)
Describe the goal of ratemaking and the difficulty in pricing insurance versus non-insurance products.
The goal of ratemaking is to assure that the fundamental insurance equation is
appropriately balanced.
Premium = Losses + LAE + UW Expenses + UW Profit
The rates should be set so that the premium is expected to cover all costs and
achieve the target u/w profit. Must consider that ratemaking is prospective
and the balance should be attained at the aggregate and individual levels.
Pricing insurance is difficult as compared to non-insurance products because
the cost (Losses+LAE+UW Expense) is unknown before the product is sold.
2. (1.75 points)
Given the following activity on five auto policies as of December 31, 2013:
Policy Effective Date
1 February 1, 2012
2 March 16, 2012
3 May 16, 2012
4 October 1, 2012
5 January 1, 2013
The exposure base is earned car years.
a. (0.5 point)
Calculate the calendar quarter earned exposure for 1q2013.
Add the months of earned exposure and divide by 12
= (1 + 0 + 3 + 2.5 + 3) / 12 = 0.792 car-years
b. (0.5 point)
Calculate the 2013 calendar year written exposure.
Policy 5 minus the cancelled exposure for policy 4.
= 0.5 - 6.5/12 = -0.042 car-years
c. (0.5 point)
Calculate the 2012 policy year written exposure.
Policies 1-4 were written in PY 2012. Subtract cancels for 2 & 4.
=4.0 - 2.5/12 - 6.5/12 = 3.25 car-years
d. (0.25 point)
Calculate the in-force exposure as of April 1, 2013.
Policies 3 & 5 are in-force
= 2 car-years
Original Expiration Date Mid-term Cancellation Date
January 31, 2013 N/A
March 15, 2013 December 31, 2012
May 15, 2013 N/A
September 30, 2013 March 15, 2013
June 30, 2013 N/A
3. (2.75 points)
A company decided to expand into the western part of a state in 2012. Written exposure growth for the total state
was 2% per month in 2012. They took a rate decrease of 10% on 1/1/2013 to increase growth. After this change,
written exposures for the state in total increased at 7% per month.
The western part of the state has on average riskier insureds with a higher average premium per exposure.
The company uses the parallelogram technique and one step trending when calculating rate indications.
a. (1.75 points)
You are completing a rate indication using data from 2012 and 2013. Fully explain the effects of the
changes on the projected premium and resulting rate indication.
The parallelogram technique assumes that the exposures are uniformly distributed over time.
The avg rel rate level for 2013 would be 1.0*50% + 0.90*50% = 0.95. However, because of the
growing exposures, more weight should be given to the 0.90 relative factor. Unless this adjustment
is made, on on-level factor, assuming uniformly distributed exposures, of 0.90/0.95 = 0.947 will be
too low, which will project on-level premiums that are too low, and thus an indication that is too high.
Premium trend could also cause problems depending on how the trend is selected. The increase
in average premium due to a higher distribution of western risks needs to be considered. The
distribution of western/non-western risks needs to be projected for the prospective period.
b. (1 point)
Briefly describe other possible changes in the experience from the new growth and the effect
on the overall indication.
High growth in general, but particularly in a new territory has potential to increase losses. New business
typically has higher loss costs. Also, the business in the west is unfamiliar to underwriters which could
could higher losses. Plus, unless there is added staff, the underwriters will not be able to spend as much
time underwriting risks, which leads to higher losses. All of these reasons would lead to higher indications
in future years. One factor that could lower future indications is the potential for expense reduction due
to the growth.
4. (2.25 points)
The following data shows the distribution of policy limits for an auto insurer over the past three
calendar years.
Written Exposures (Car-Years)
Calendar 100,000 250,000 500,000
Year Limit Limit Limit
2010 3,000 4,500 2,500
2011 2,100 5,500 3,400
2012 1,000 6,700 4,400
a. (0.5 point)
Given the following increased limits factors, suggest a premium trend rate for shifts in policy limits.
100,000 250,000 500,000
Limit Limit Limit
ILF 1.00 1.10 1.20
Avg ILF Trend
2010 1.095
2011 1.112 1.0154
2012 1.128 1.0146 Select trend of 1.5%
b. (0.25 point)
State an assumption made if the actuary chooses to use the selected trend factor.
Assumes the shift in increased limits observed in the experience period will
continue in the future.
c. (0.5 point)
Describe two other changing factors that contribute to premium trend.
Changes in deductible - if insureds are increasing their average deductible
over time, the average premium will have a downward drift
Changes in mix of business - growing in new classes/territories can
gradually change the average premium over time.
Other possible asnwers - model year/vehicle symbol, changing discounts
Note - you can use your experience here to answer this if you don't
remember the answers from the book.
d. (1 point)
Explain why determining premium trend by analyzing the shift in individual rating variables
may not be practical and describe an alternative technique.
The needed data may not be readily available and there may be many
variables causing premium shifts that are relatively small and difficult
to quantify. Also, you may not be able to identify all of the individual
variables that contribute to premium drift.
Measuring all premium shifts simultaneously is ofter a better technique
as it picks up all variables. To do this, we look at the historical average
premium per exposure, adjusted for all one-time shifts, such as rate
changes.
5. (2.5 points)
A law change effective 5/15/2013 is estimated to increase benefits to insureds by 10% on average.
a. (0.5 point)
The law change affects policies written on or after 5/15/2013. Calculate the adjustment factor to bring 3rd quarter
losses to current benefit level. You are doing an accident year analysis.
Note - you'll want to draw diagrams for a. - c.
Adjustment = 1.10 / (1.00 * 3/4 + 1.10 * 1/4) = 1.0732
b. (0.5 point)
The law change affects policies written on or after 5/15/2013. Calculate the adjustment factor to bring 2nd quarter
losses to current benefit level. You are doing a policy year analysis.
Adjustment = 1.10 / (1.00 * 1/2 + 1.10 * 1/2) = 1.0476
c. (0.5 point)
The law change affects losses occurring on or after 5/15/2013. Calculate the adjustment factor to bring 2nd quarter
losses to current benefit level. You are doing a policy year analysis.
Adjustment = 1.10 / (1.00 * 1/32 + 1.10 * 31/32) = 1.0028
d. (0.75 point)
Explain why an actuary may divide the experience into quarters to bring losses to current benefit level.
The adjustment uses the parallelogram method which assumes a uniform distribution losses. This
assumption may not be appropriate for losses that are affected by seasonality. Therefore, losses
broken down into quarterly detail are used.
6. (2.5 points)
You are pricing personal automobile bodily injury coverage in State X. You have annual basic limits
loss & ALAE data for State X available. Given the following information, select trends to use for
adjusting the losses to calculate the indicated change. The indication methodology uses five years
of company data. Fully explain your selections.
State X Basic Limits Ultimate
Claims
Frequency Severity
2005 0.0830 $2,100
2006 0.0820 $2,600
2007 0.0680 $2,800
2008 0.0630 $3,100
2009 0.0650 $3,300
2010 0.0650 $3,100
2011 0.0700 $3,300
2012 0.0680 $3,000
Annual Exponential Fit
# of Frequency Severity
Points Fit
R
2 Fit
R
2
8 -2.7% 42.1% 4.9% 61.6%
6 0.9% 18.8% 1.4% 15.6%
4 2.1% 54.4% -2.2% 36.3%
Company Countrywide Selected Trends
Frequency 0.5%
Severity at Basic Limits 3.0%
Severity at Total Limits 1.5%
Company State X Selected Trend for Reserve Analysis
Severity at Total Limits 2.0%
Trends Selected Using Industry Data for State X
Frequency 1.0%
Severity at Basic Limits 2.0%
Severity at Total Limits 4.0%
There are several things to consider here
1. State vs. Countrywide vs. Industry Data
2. Comparing B/L and Total Limits selections
3. R-squared
There is no one right answer here - here is a sample
Many differences in the trends of losses make the use of State X for the insurer the ideal
data to use for selections. These differences include different types of insureds between
states and insurers, differences in laws and regulations, and differences in operating
procedures. However, the low goodness of fit for the company state X data forces us
to consider this data.
Accident Year
For frequency, the 4 point fit has the best fit. Prior to 2008, we see a significant decrease
in frequency. This could be due to many factors, including a shift in type of insured.
The average of the 4 & 6 point fit is 1.5%. This is higher than the company countrywide
and the state X industry trend. I would judgmentally give 50% weight to the industry
trend and 25% to the fitted 4 & 6 point trends each. Selected 1.25%.
For severity, we want the B/L trend to be less than total limits. We see that the B/L trend
for both company countrywide and state X industry is half of the countrywide trend.
The 4 & 6 point fits have lower r-squared than the frequency fits. When we look at the
last five years, there isn't an obvious pattern. You could justify a selection of 0%.
However, using the B/L and total limits trends for countrywide and industry, we will
select 50% of the selected company state X trend of 2%. Selected 1%.
7. (2.5 points)
You are given the following Homeowners information:
Written Premium $9,500,000
Earned Premium 9,000,000
Written Exposures (House Years) 12,000
Earned Exposures (House Years) 11,500
Commissions 1,200,000
General Expenses 990,000
Other Acquisition Expenses 540,000
Licenses and Fees 55,000
Taxes 275,000
Projected Loss & LAE Ratio 64.0%
Profit and Contingencies Prov 2.0%
Employee Cost Index 3.0%
Consumer Price Index 1.0%
% of Gen and Oth Acq used for
Employee Cost 50.0%
Trend Period 2 years
75% of General and Other Acquisition expenses are assumed to be fixed.
100% of Licenses and Fees are assumed to be fixed.
Projected Pure Premium = $600
a. (1.5 points)
Calculate the indicated rate per exposure.
Fixed Variable
Fixed Ratio* Variable Ratio
General $742,500 64.57 $247,500 2.75%
Oth Acq $405,000 33.75 $135,000 1.42%
Lic&Fees $55,000 4.58 $0 0.00%
Comm $0 0.00 $1,200,000 12.63%
Tax $0 0.00 $275,000 2.89%
Total 102.90 19.70%
Selected Fixed Expense Trend = 3% * 50% + 1% * 50% = 2%
Projected Fixed Expense per Exposure = 102.9 * 1.02
2
= 107.06
Ind Rate = 600 + 107.06 = $902.98
1 - 19.7% - 2%
b. (0.5 point)
Briefly describe two shortcomings of the Exposure-Based Expense Projection Method.
1. Requires the actuary to split expenses into fixed and variable
2. Proposed procedure allocates countrywide fixed expenses to each state based on exposures
Average fixed expense levels may vary by location (e.g., advertising costs)
3. Some expenses considered fixed vary by other characteristics
E.g., New vs. renewal business - only affects if distribution changing significantly
and/or varies significantly by state
4. Existence of economies of scale in a changing book will lead to increasing or decreasing
projected average expense per exposure figures
c. (0.5 point)
Briefly describe a situation when trending the fixed expenses is not necessary.
Potential answers
1. Use the premium-based projection method - if the average fixed expenses and average
premium are changing at the same rate, then the fixed expense ratio will be consistent
and no trending is necessary
2. If an inflation-sensitive exposure based (e.g., AOI) is used, then no trending is necessary
if the expenses and exposure base are changing at the same rate.
8. (3.5 points)
The proposed effective date for the revised workers compensation advisory loss costs is January 1, 2014.
You are given the following information.
(1) (2) (3) (4)
Reported Med Loss Medical Fee Portion of Medical
Accident Med Losses Development Schedule Losses Subject
Year Ending ($000) Factor to Ult Change to Fee Schedule
June 30, 2010 1,600,000 1.400 0% 75.0%
June 30, 2011 1,100,000 1.600 -30% 70.0%
June 30, 2012 1,000,000 1.900 0% 70.0%
June 30, 2013 700,000 2.900 10% 70.0%
Projected Loss Cost Premium for AYs 7/1/2009 - 6/30/2013 15,000,000
Projected Future Medical Fee Schedule Change 0.0%
Projected Portion of Medical Losses Subject to Fee Schedule 70.0%
Current and Projected Annual "Other Medical" Level Change 3.0%
Expected Indemnity Loss Ratio 34%
Expected LAE Ratio to Ultimate Loss 20%
a. (2.5 points)
Calculate the indicated change to the workers compensation advisory loss costs, including
LAE given the following information and your answers from a. and b.
(5) (6) (7) (8) (9) (10)
Combined Factor to Combined Fact to Adj Projected Expected
Accident Effect of Cur Medical Effect of Med Ben to Ult Med Med Loss
Year Ending Med Trends Cost Level Proj Trend Proj Cst Lv Losses Ratio
June 30, 2010 0.8% 0.870 1.018 0.886 1,984,640
June 30, 2011 -20.1% 1.089 1.018 1.109 1,951,840
June 30, 2012 0.9% 1.079 1.018 1.099 2,088,100
June 30, 2013 7.9% 1.000 1.018 1.018 2,066,540
8,091,120 53.9%
(5)= (3)*(4) + (1.0-(4))*3%
(6)=[1.0 + (5NextRow)] * (6NextRow)
(7)=(1+3%)^2 * 30% + 70% *note projected trend period 1/1/13 - 1/1/15
(8)=(6) * (7)
(9)=(1) * (2) * (8)
(10)=(9tot) / 15,000,000
Expected Loss & LAE Ratio = (53.9% + 34%) * (1 + 20%) = 105.5%
Indication = 5.5%
b. (1 point)
Insurer X plans to adopt the 1/1/2014 advisory loss costs. Given the following company information,
calculate the overall company rate change.
General Expenses 11.0%
Other Acquisition Costs 7.0%
Taxes, License and Fees 2.5%
Commisions and Brokerage Fees 9.0%
Target Profit Provision 0.5%
Expected Loss Cost Difference -5.0%
Current Deviation 1.400
(1) Total Expense and Profit 30%
(2) Expense and Profit Adjustment 1.429
= 1.0 / [1.0 - (1)]
(3) Operational Adjustment 0.95
= 1.0 - 5%
(4) Proposed Deviation 1.358
= (2) * (3)
(5) Current Deviation 1.400
(6) Industry Loss Cost Change (from 2.c.) 5.5%
(7) Company Change 2.4%
= (4) / (5) * [1.0 + (6)] - 1.0
9. (3 points)
The three primary purposes of risk classification are
To protect a program's financial soundness
Be fair
Permit economic incentives to operate and thus encourage widespread availability of coverage
a. (1.75 points)
Describe the difficulty of fairness in risk classification.
LOTS of potential answers here. This is how I would answer this.
Fair in risk classification interchangeable with equitable. Differences in prices among classes should
reflect differences in expected costs. There should be no subsidy among the classes. Risks within
a class should also have the same expected costs. If a risk has a substantially different expected cost,
it should be moved to a different class. It is difficult to identify the risk characteristics and establish
classes that do this. Another difficulty comes from the social aspect of fairness. If grouped accurately,
some classes will be grouped by characteristics or at level deemed socially unacceptable. It is challenging
to find the balance between individual equity and social adequacy.
b. (0.5 point)
Explain whether risk classification predicts the costs for an individual risk in a class.
Risk classification is not intended to do this, but instead predict the expected costs for a
particular class of risks. It is impossible and unnecessary to predict costs for any
individual risk.
c. (0.75 point)
To achieve the primary purposes, briefly describe three basic principles that should be present in any
sound risk classification system.
Should reflect expected cost differences
Should distinguish among risks on the basis of relevant cost-related factors
Should be applied objectively
Should be practical and cost-effective
Should be acceptable to the public
10. (1 point)
Briefly describe two challenges in territorial ratemaking.
1. Tends to be heavily correlated with other rating variables
E.g., high value homes often located together
Makes traditional univariate analysis very susceptible to distortions
2. Often analyze territory as collection of small units
Data in each individual territory is sparse
11. (1 point)
Describe the primary shortcoming of univariate approaches. Give an example
Do not accurately take into account the effect of other rating variables.
Example - a one-way analysis may show in personal auto that older
cars have high claims experience relative to newer cars. However, this
is distorted because older cars tend to be driven by younger drivers
with worse claims experience.
12. (1.25 points)
Describe how the LER approach does not recognize behavior differences of insureds and the effect
on the calculated deductible credits. Discuss one way to recognize these behavior differences.
1. Method assumes insureds at different deductible levels have same claiming behavior.
In reality insured with $250 deductible is more likely to make a claim for $1,100 than an
insured with a $1,000 deductible
2. Lower-risk insureds tend to choose higher-deductibles because they realize they're less
likely to have a claim.
LER approach does not recognize these behavior differences, so higher deductible policies
may end up being more profitable.
GLMs do not assume frequency is same for all risks. The GLM results are influenced
by both the limiting of losses and behavioral differences among insureds at different
limits.
13. (2.5 points)
Given the following information:
Property Value = $250,000
Amount of Insurance = $175,000
Coinsurance Requirement = 80%
a. (1 point)
Calculate the coinsurance penalty for the following losses.
i. $90,000
ii. $190,000
iii. $220,000
Coinsurance requirment = $250,000 * 80% = $200,000
Coinsurance Penalty % = $175,000 / $200,000 = 87.5%
i. $90,000 - $90,000 * 87.5% = $11,250
ii. $175,000 - $190,000 * 87.5% = $8,750
iii. No penalty because L > coinsurance requirement
b. (0.5 point)
Identify the problem with underinsurance from the insurer's perspective.
The pure premium rate decreases as the policy face increases. If policyholders are
underinsured, this is a problem from insurers perspective because if rates are
calculated assuming all properties are insured to value, the premium charged will
not be adequate to cover expected losses arising from those policies not insured to value.
c. (0.75 point)
An insurance company increases the insurance to value of its book of business. Briefly describe
the impact on each of the following:
Premium
Losses
Expenses
Premium could see higher prem. as a result of larger exposure amounts written
! Could see lower premium if there are higher cancel/non!renews
Losses expect to see larger total and near total claim amts from larger exposures
! Losses may decrease from higher cancel/non!renew
! Losses may decrease if reinspection also leads to loss control measures
implemented by homeowners.
Expenses increased inspection/reinspection may create additional expenses, however
increase relative to premium change is unclear.
d. (0.25 point)
Identify another way for insurers to handle varying levels of insurance to value, other than
coinsurance penalties.
Base the rates charged on the level of insurance to value. The higher the ITV, the lower
the rate.
14. (2.5 points)
Using the ISO experience rating plan for a policy with premises/operations coverage and the
following information, calculate the experience debit or credit. Show all work.
Expected Percent of
Basic Limits
Detrend Loss & ALAE Unreported
Policy Period Factors as of September 30, 2013
2010 0.864 10%
2011 0.907 25%
2012 0.952 40%
Policy being rated will be in effect January 1, 2014 - December 31, 2014.
Premises/operations premium is $400,000.
Reported loss and ALAE for experience period as of September 30, 2013 (limited
by basic limits losses and MSL) is $425,000.
Expected experience ratio is 0.92
Expected loss and ALAE ratio is 0.67
Maximum single limit per occurrence is $100,000.
Credibility is 70%
All policies in experience period are occurrence policies.
(1)=0.67*$400,000 (3)=(1)*(2) (5) (6)=(3)*(4)*(5)
Cur Comp Company (4) Exp % Exp
B/L Loss (2) Subject Expected B/L Loss B/L Loss
Policy &ALAE Detrend B/L Loss Experi &ALAE &ALAE
Period Costs Factors &ALAE Ratio Unrptd Unrptd
2010 268,000 0.864 231,552 0.92 10% 21,303
2011 268,000 0.907 243,076 0.92 25% 55,907
2012 268,000 0.952 255,136 0.92 40% 93,890
729,764 171,100
(7) Rpt Loss 425,000
(8) Exp Unrptd 171,100 =(6)
(9) Exp Ultimate 596,100 =(7)+(8)
(10) Comp Subj B/L 729,764 =(3)
(11) Actual Exp Ratio 0.8168 =(9)/(10)
(12) Exp Exper Ratio 0.9200
(13) Credibility 0.7000
(14) Exp (Credit)/Debit -0.0785 =[(11) - (12)] / (12) ! (13)
15. (2 points)
The fourth principle of Property and Casulaty Insurance Ratemaking state:
A rate is reasonable and not excessive, inadequate, or unfairly discriminatory if it is an
actuarially sound estimate of the expected value of all future costs assocated with an
individual risk transfer.
a. (1 point)
Defend the assertion that experience rating supports the principle that a rate should not
be unfairly discriminatory.
This problem was taken from Exam 8, but thought it could be answered from the Exam 5 material.
The main goal of experience rating is individual risk equity. Experience rating
recognizes that each risk has a different loss potential, so by modifying rate
appropriately, the expected profit potential for each risk can be made equal. This
ensures that equity is achieved and rates are not unfairly discriminatory.
Otherwise, risks with lower loss potential within the class would subsidize other
higher loss potential risks.
b. (1 point)
Suppose the industry experience rating plan assigns too much credibility to individual experience
for large insureds and assigns too little credibility to individual experience for small insureds. Argue
that in a competitive insurance market, rates will not be unfairly discriminatory.
Large risks with credit mods will have worse experience than expected. Similarly,
small risks with credit will have better experience that expected. Forces of supply
and demand in competitive market will push rates down for large debit and small
credit risks. All risks now pay rates commensurate for their exposure to loss so
rates are not unfairly discriminatory
16. (2.5 points)
You are given the following information:
Class A Class B
Premium-First Year $867.84 X
Loss Cost-First Year $700 $500
Fixed Expense
First Year $100 $90
Subsequent Years $50 $40
Variable Expense
First Year 10% 10%
Subsequent Years 5% 5%
Persistency Rate 80% 90%
Average annual loss cost trend is 3%.
Average annual premium trend is 2%
There is no expense trend.
Interest rate for discount is 5%
Premium-to-Surplus ration is 2:1
Target pre-tax return on equity is 8%
Loss costs for renewal business are 5% lower than for new business.
Calculate the indicated rate relativity for Class B as compared to the base class (Class A) using the
asset share pricing model and a two-year time horizon.
PV Variable Expense Fixed Expense
Policy Premium Loss Year 1 Renewal Year 1 Renewal
Year (1) (2) (3) (4) (5) (6)
1 X 500.00 0.10X 0.00 90.00 0.00
2 1.02X 489.25 0.00 0.051X 0.00 40.00
Policy Persistency Discount Present Value
Year Rate Cumulative Profit Factor Profit Premium
(7) (8) (9) (10) (11) (12)
1 1.000 1.000 0.9X-590 1.000 0.9X-590 X
2 0.900 0.900 .8721X-476.3 1.050 .831X-453.64 0.8743X
!"#$%
!
1.731X-1043.64 1.8743X
We want return on premium to equal 8% / 2 = 4%
(1.731X - 1043.64) / 1.8743X = 0.04
! X = 630.21
Indicated rate relativity = 630.21 / 867.84 = 0.726
17. (3 points)
a. (1.5 points)
Fully explain three examples of how a change in the mix of business can affect a reserve analysis.
Suggest a way to handle each change.
1. Changes in portfolio volumes - if there is a high growth rate, it can cause distortions in
development factors due to significant shifts in the average accident date in the exposure
period. May substitute accident quarter for accident year data.
2. Change in policy conditions, such as deductibles, policy limits - this can change the
size of loss distribution, changing the development. May substitute policy year data
for accident year data.
3. Change in type of business written - this can include change in classes or territories.
The different business may have different development patterns or loss ratios. One
option would be to split the data into more homogeneous groups.
b. (1.5 points)
Fully explain three additional factors that can affect claim settlement patterns and the
resulting reserve analysis.
1. Changes in settlement procedures, such as change in case outstanding strength or claims
closure rate. These cause changes in development patterns. One option is to use
Berquist/Sherman adjustments.
2. Jury Awards / Legislation / Regulation - these can change type of loss, size of loss, and/or
reporting patterns. May need to substitute report year data for accident year because
loss may more closely correlate with report date.
3. Large Claims - individual large claims can distort reporting and/or payment patterns. One
option is exclude these from the experience and evaluate separately.
18. (2.5 points)
Project the ultimate 2013 accident year losses given the following data. Discuss any patterns in the data and
any reasons for the pattern.
Accident Auto Collision Paid Claims
Half-year 6 12 18 24 30 36
2010-1 7,500 9,000 9,100 9,110 9,110 9,110
2010-2 8,000 10,800 10,925 10,940 10,940 10,940
2011-1 9,000 10,800 10,900 10,910 10,910 10,910
2011-2 8,400 11,350 11,500 11,510 11,510
2012-1 9,450 11,300 11,400 11,415
2012-2 8,500 11,500 11,575
2013-1 9,925 11,900
2013-2 9,000
6-12 12-18 18-24 24-Ult
2010-1 1.200 1.011 1.001 1.000
2010-2 1.350 1.012 1.001 1.000
2011-1 1.200 1.009 1.001 1.000
2011-2 1.351 1.013 1.001 1.000
2012-1 1.196 1.009 1.001
2012-2 1.353 1.007
2013-1 1.199
First Half 1.199 1.010 1.001 1.000
Second Half 1.351 1.010 1.001 1.000
Selected 1.351 1.010 1.001 1.000
6-Ult 12-Ult 18-Ult 24-Ult
CDF 1.366 1.011 1.001 1.000
Ult Paid 12,294 12,031
Total AY 2013 24,325
Seasonality is present in the data. For auto collision, more accidents may occur in the
winter months. These increase claims in December have less time to be paid before year end.
Claims in January and February have until June 30 to be paid. These accidents in December
being paid after year end increase the 6-12 month development factor.
19. (2.75 points)
You are given the following information:
Reported Claims
Accident Earned (Age of Development in Months)
Year Premium 12 24 36 48
2010 $55,000 $12,000 21,000 26,250 28,875
2011 56,000 13,000 22,750 28,438
2012 57,000 14,000 24,500
2013 58,000 15,000
Selected 48-Ultimate Reported Development Factor = 1.10
Selected Ultimate Claims Ratio = 68%
a. (1.5 points)
Calculate the indicated IBNR using the Bornhuetter-Ferguson and Benktander techniques.
Accident Initial Expected Age-to-Age Factors
Year Ultimate Claims 12-24 24-36 36-48 48-Ult
2010 $37,400 1.75 1.25 1.10 1.10
2011 $38,080 1.75 1.25
2012 $38,760 1.75
2013 $39,440
Selected 1.750 1.250 1.100 1.100
CDF 2.647 1.513 1.210 1.100
AY BF Ult* Benk Ult** BF IBNR Benk IBNR
2010 32,275 31,809 3,400 2,934
2011 35,046 34,520 6,609 6,082
2012 37,634 37,252 13,134 12,752
2013 39,539 39,601 24,539 24,601
*BF Ult = Rpt + Initial Exp Ult * (1-1/CDF)
*Benk Ult = Rpt + BF Ult Ult * (1-1/CDF)
b. (1.25 points)
Analyze the results from a. and discuss the reasonableness of using these techniques for selected IBNR for
this data.
Benktander
Accident Ultimate Claims BF indicates a higher IBNR for AYs 2010-2012. This is
Year Ratio due to the initial expected claims ratio of 68%. Looking
2010 57.8% at the Ultimate Claims ratios, it appears this is too high
2011 61.6% for older accident years. The data shows a deteriorating
2012 65.4% loss ratio. Because development is consistent, the rept
2013 68.3% development technique would be a better option. Since
Benktander is closer to the development estimate, so it
would be a better selection than the BF.
20. (2.25 points)
Fully explain the value in using frequency-severity techniques. Include in the discussion advantages compared to
development techniques.
Development techniques can be unstable and innaccurate in the most recent accident years. Freq-Sev
techniques are able to break the claims down into the freq and sev components. Claims counts
tend to be stable and good estimates of ultimates can be made using the development technique.
Severities from older accident years can be accurately estimated then using assumptions on the losses,
such as inflation, the severities for more recent accident years can be established. Being able to explicitly
reflect inflation in the projection instead of it being included in the selection of development factors
is an advantage of the method. Other advantages are the potential to gain greater insight into the
claims process, such as claims reporting, settlement, and avg cost of claims. This helps us to continue
to ask the right questions and allow us to find the best estimate of ultimate claims. Also, freq-sev
tech can be used with paid data only so it is independent of case outstanding changes.
21. (3.25 points)
Given the following loss information as of 12 months maturity for accident years 2010
through 2013:
Reported Closed
Accident Paid Claims Claims Claim Open Claim
Year ($000) ($000) Counts Counts
2010 22,200 43,000 4,900 2,660
2011 36,400 90,000 7,800 6,200
2012 55,000 122,500 11,450 7,000
2013 85,500 190,000 17,250 9,500
a. (1.5 points)
Test the above data for changes in case reserve adequacy and interpret the results.
Case Average Annual Average Annual
AY Outstanding Open Claim Change Paid Change
2010 20,800 7,819.55 4,530.61
2011 53,600 8,645.16 10.6% 4,666.67 3.0%
2012 67,500 9,642.86 11.5% 4,803.49 2.9%
2013 104,500 11,000.00 14.1% 4,956.52 3.2%
When we compare the annual rates of change between average case outstanding
and average paid claims, we observe the average increase in average case
outstanding is much higher. These higher trends are indicative of changes in
case outstanding adequacy (i.e., case reserve strengthening).
b. (0.75 point)
Describe the leveraging effect that a change in case reserve adequacy has on the
IBNR indicated by the reported loss development method.
With case outstanding strengthening, the reported claim development factors
will increase. These higher development factors will be applied to reported
claims that are now at a higher level. The result is an estimate of unpaid claims
that will be overstated.
With case outstanding weakening, the opposite is true. Smaller development
factors applied to smaller reported claims, and thus an understated unpaid
claim estimate.
c. (1 point)
Use the Berquist-Sherman technique for case reserve adequacy to calculate the
adjusted reported claims for each accident year.
Based on the paid claim trend, select a trend rate of 3%. Use the most recent
case reserve adequacy, and trend back using the selected 3%.
Adj Avg Adj Case Adj Rpt Clm
AY Open Claim ($000) ($000)
2010 10,066.56 26,777 48,977
2011 10,368.56 64,285 100,685
2012 10,679.61 74,757 129,757
2013 11,000.00 104,500 190,000
22. (2 points)
Given the following information as of December 31, 2013:
Paid Development
Dev Factor Rato of Factor to
Accident Paid Claims to Ultimate Received S&S Ultimate for
Year Gross of S&S Gross of S&S to Paid Claims S&S Ratio
2011 $27,654 1.075 0.35 1.000
2012 $24,388 1.150 0.37 1.010
2013 $14,680 1.750 0.28 1.300
a. (1.5 pts)
Use the ratio method to estimate the recoverables for salvage and subrogation (S&S) for accident
years 2011 - 2013.
[1] [2] [3] [4] [5] [6] [7]
Accident Received S&S Dev. Factor Paid Dev
Year Paid Claims to Ult Ult Ratio Paid Claims Factor to Ult Ult Claims
2011 0.35 1.000 0.3500 $27,654 1.075 29,728
2012 0.37 1.010 0.3737 $24,388 1.15 28,046
2013 0.28 1.300 0.3640 $14,680 1.75 25,690
[10]
[8] [9] S&S
Ult S&S Paid S&S Recoverables
10,405 9,679 726
10,481 9,024 1,457
9,351 4,110 5,241
7,424
[4] = [2] x [3] [9] = [2] x [5]
[7] = [5] x [6] [10] = [9] - [8]
[8] = [4] x [7]
b. (0.5 pt)
Briefly discuss the advantages of using the ratio method to determine salvage and subrogation recoverables.
The development factors for the ratio method are less leveraged at early maturities than
development factors would be in the reported recoveries or received recoveries
development methods. The ratio method provides Ult S&S ratios to paid claims, which can be
used as a diagnostic, so that the actuary may use judgement in selecting a more reasonable
S&S ratio for AY's that show odd behavior.
23. (3 points)
Given the following to be used in a reserve review:
Reported Reported
Accident Claims Claims Development Factors to Ult
Year @12/31/2012 @12/31/2013 Age (Months) Method 1 Method 2
2010 $280,000 $310,000 12 3.20 2.90
2011 175,000 220,000 24 1.35 1.46
2012 100,000 220,000 36 1.12 1.16
2013 140,000 48 1.05 1.05
a. (1.5 points)
There are only two methods used for selecting ultimate claims. Retrospectively test thest methods and
analyze the results.
Estimated IBNR Emergence
in CY 2013 Actual Difference
Acc Year Method 1 Method 2 Emergence Method 1 Method 2
2010 $18,667 $29,333 $30,000 -11,333 -667
2011 35,938 45,259 45,000 -9,063 259
2012 137,037 98,630 120,000 17,037 -21,370
Total 191,641 173,222 195,000 -3,359 -21,778
Method 2 worked really well for losses at 24 and 36 months. It significantly underestimated IBNR
emergence for losses at 12 months.
Method 1 worked well in total. However, it significantly underestimated IBNR emergence
at 24 and 36 months and significantly overestimated IBNR emergence at 12 months.
b. (0.75 point)
Estimate the IBNR emergence for accident year 2013 in calendar year 2014. Describe your method.
If we select Method 2 for 24-ult, we can calculate the 12-24 for method 1 to be 2.19. Using
this, we would get IBNR emergence of 119,000 for AY 2013 in CY 2013, which is very close
to the actual 120,000.
Est IBNR Emergence for AY 2013 in CY 2014 = $140,000 * (2.19 - 1.0) = $166,000
c. (0.75 point)
Describe the accuracy of using the given development factors for projecting the expected emerged
claims for accident year 2013 in the first quarter of 2014.
We would need to make an assumption on how claims emerge throughout the year. A common
assumption is to use linear interpolation. However because development tends to slow down
overtime, more losses should be reported during the first quarter of the year. This would result
in a 15-ult factor which is smaller than linear interpolation would suggest, and therefore
understate 1st quarter emergence.
24. (2.25 points)
Berquist and Sherman recommend that wherever possible, an actuary should include the concepts
of credibility, regression analysis, and data smoothing in the actuarial reserve analysis. Describe
three examples of how an actuary can incorporate these.
Selecting age-to-age and tail factors - credibility and data smoothing are considered
using several years of data and looking at different averages for selecting factors.
Often, for tail factors, curves are fit to development factors and extrapolate.
Credibility is often not explicitly calculated when selecting factor, but reflected
implicitly. Actual data may also be judgmentally weighted with industry factors.
Trend factors used when analyzing frequency, severities and pure premium - regression
analysis is used often for selecting these trends. We also may credibility
weight these company trends with industry data. Again, credibility and data smoothing
are considered as we consider the number of years and weights to more recent
experience.
Selection of claim ratios - same explanation as trend
Determing adjustments for operational changes - for our Berquist-Sherman adjustments,
we use regression analysis to select trends and estimate paid claims for a given
claims closure rate.
Note - another approach to this problem would be to first address credibility, regression,
and data smoothing - describe the general ways they are incorporated, then fit them
to the examples.
25. (2.75 points)
An insurer currently combines the experience of personal auto property damage and commerical
auto property damage when estimating IBNR. Historically, we have the following data, which has
been in a steady-state for the past five years.
Ultimate Percentage Reported Annual
Claims Ratio 12 months 24 months 36 months Growth
Personal 75% 70% 90% 99% 5%
Commercial 65% 55% 80% 95% 5%
Total 70%
a. (2.25 points)
Evaluate the effect of a change in growth rate of commercial auto property damage to 20%
on the reported development, Bornhuetter-Ferguson, and Cape Cod techniques.
First, the overall claim ratio of the grouped business will decrease. Without an explicit
change to the expected loss ratio, the BF method will overstate expected losses.
The cape cod, which uses actual experience to determine expected claims will be
more responsive than BF, but will still overestimate initial expected.
Second, commercial has a longer reporting pattern, and therefore higher development
factors. As the proportion of commercial increases, the age-to-age factors will increase,
but because we usually rely on an average of several years, the selected factors will
not be responsive enough for the changing conditions. The result is CDFs that are
not large enough. This will cause the development method to understate IBNR.
Combined - Reported Development Technique will understate due to the
insufficient development factors. The excessive claim ratio will
have no effect
BF will overstate initial expected, but lower development factors
will understate the % unreported. These are opposing forces so
the estimate could be higher or lower. Cape Cod will do the same
except is more responsive to the loss ratio change, so expected
losses will not be as overstated.
a. (0.5 point)
Describe the expected effect of a deteriorating loss ratio for all business combined in addition
to the changing mix of business.
Because the development factors are understated, and now the expected losses will
likely be understated due to the loss deterioration, all methods will understate the IBNR.
26. (2 points)
You are given the following information for a small book of business.
Calendar Earned Paid Paid
Year Premium ULAE Claims
2010 3,000,000 75,000 440,000
2011 4,000,000 100,000 120,000
2012 6,000,000 180,000 2,500,000
2013 10,000,000 275,000 1,500,000
Expected Claims Ratio for each Accident Year = 60%
Claims for an AY are expected to be paid out 20% per year
Case Outstanding at 12/31/2013 = $2 million
Total IBNR at 12/31/2013 = $3.52 million
Pure IBNR at 12/31/2013 = $3 million
a. (1.75 points)
Calculate the estimated unpaid ULAE at 12/31/2013 using the Mango-Allen Refinement Technique
Accident Expected
Year Claims 2010 2011 2012 2013
2010 1,800,000 360,000 360,000 360,000 360,000
2011 2,400,000 480,000 480,000 480,000
2012 3,600,000 720,000 720,000
2013 6,000,000 1,200,000
Total 360,000 840,000 1,560,000 2,760,000
ULAE Ratio
Calendar Paid Expected Paid ULAE-to-
Year ULAE Paid Claims Expected Pd
2010 75,000 360,000 0.208
2011 100,000 840,000 0.119
2012 180,000 1,560,000 0.115
2013 275,000 2,760,000 0.100
Total 630,000 5,520,000 0.114
Selected ULAE Ratio 0.110
Case Outstanding at 12/31/2013 2,000,000
Total IBNR at 12/31/2010 3,520,000
Pure IBNR at 12/31/2010 3,000,000
Estimated Unpaid ULAE at 12/31/2010 Using Pure IBNR 468,600
b. (0.25 point)
Explain whether this technique is appropriate for this data.
Yes because it is used for books of business with a small amount of widely
varying claims.
Expected Claims Paid in Calendar Year