7.

IBNR
1. Part
2. Part
3. Part

1
OUTLINES OF THIS SECTION

In Part 1, after the definition of terms, we are going to

introduce the following Reserve Estimate Methods:
i. Payed Loss development Method (PLDM)
ii. Incurred Loss development Method (ILDM)

In part2, we are going to compare and consider two methods

introduced in part1.

In part3, we are going to introduce methods other than PLDM

and ILDM, mainly Bornhuetter-Ferguson Method, and also
briefly explain prediction of Loss Survey Cost.

2
PART

3
CONTENTS IN PART 1

1. Definition

2. Points to be considered

3. Basic Reserve Estimate Methods

- Payed Loss Development Method (PLDM)
- Incurred Loss Development Method (ILDM)

4
DEFINITION

What is the Loss Reserves?

The amount necessary to pay out the Unpaid Insurance Payment

Why is the Loss Reserves important?

Accurate measurement of financial situation and revenue

5
DEFINITION (CONTINUED)

Accounting aspect of Loss Reserves

Balance Sheet

Assert Debt

Surplus

6
DEFINITION (CONTINUED)

Carried Loss Reserves

The amount on financial statements created for external and internal use

Indicated Loss Reserves

The amount of loss reserves estimated by applying some methodology

Reserve Margin/Deficit
The difference between the Carried Loss Reserves and the Valuation Reserves

7
DEFINITION(CONTINUED)

Terms Related to Loss Reserves

- IBNR Reserves [Incurred But Not Reported Reserves

- Receiving in Process Claim
- Loss Reserves and Calculation Formulas of Loss Reserves
- Increase of Existing Claim
- Payment Amount Not Settled Loss Reserves

8
DEFINITION(CONTINUED)

Case Loss Reserves

- Reported Accrued Claims
- Amount Estimated by Adjuster or Formula

IBNR Reserves in a Broad Sense

- Pure IBNR Reserves
- Receiving in Process Claim
- Increase of Existing Claim
- Payment Amount Not Settled Loss Reserves

9
 DEFINITION(CONTINUED)

Damage Research Expenses are divided into the following two costs:

Counsel's expenses Cost reduction costs (DCC)

- Including all defense, lawsuits, medical expenses control fee, regardless of internal
or external
- Generally, the cost required to control the damage scale of each insurance accident

Assessment Other Expenses (AO)

- ALL Including the Complaint Assessment costs, regardless of internal and external
- Generally, including ALL the Costs associated with recording and assessing insurance
accidents

10
 LIFE CYCLE OF LOSS RESERVES
7/11/01 8/1/01
Accident report As a result of the calculation, the
result of the Reserves Estimate $
Receiving in process 1000 is entered as Loss. reserves
4/2/01 complaint
Accident occurred

8/18/02
1/1/02 10/5/01
Payment insurance fixed
Reassessment Individual Loss Reserves
Calculation
$ 30,000 Loss reserves
$ 30,000 Loss reserves
$ 10,000 Loss Reserves

8/25/02 9/2/02
Send Insurance Payment Data
Insurance Payment
$ 30,000 Loss Reserves
Close

11
DEFINITION (CONTINUED)

Reserves = Liabilities = Accruals = Unpaid

= Case Reserves + IBNR Reserves

Incurred Losses and Claim counts may have various meanings!

- Final Losses (including IBNR Reserves)
- Report Based Losses (excluding IBNR Reserves)

12
ITEMS TO BE CONSIDERED:
- DATA

Accident Occurrence Date

The Date when the insurance accident occurred

Accident Report Date

The Date when the insurance accident was reported to the insurance
company for the first time

Accident Record Date

The Date when the insurance accident was recorded as the statistical record of
the insurance company for the first time

13
ITEMS TO BE CONSIDERED:
- HOMOGENEITY
Accuracy is often improved by dividing claims into groups with
similar characteristics and taking up experience values.

Car Insurance

Liability Physical Damage

Interpersonal
Objective
Medical Bills
Non-Insurance Car Correspondence - Collision
Interpersonal Other than Collision
For Non-Insurance Vehicles -
Objective 14
ITEMS TO BE CONSIDERED:
- RELIABILITY

The group of claims should be large enough to be statistically

reliable
Decompose a group small while maintaining reliability of statistical data

Use of auxiliary data

For example, industry data or nationwide data

15
ITEMS TO BE CONSIDERED:
- APPEARANCE PATTERN

Difference between the occurrence of insurance accident and

the record as accident record by insurance company

- Objective claims are reported promptly

- The indemnity liability is delayed greatly

16
ITEMS TO BE CONSIDERED:
- PAYMENT PATTERN

The length of time normally taken until the claimed insurance

amount is paid up (or closed)

It affects the method of calculating Loss Reserves

- The product line to which the Insurance is promptly paid does not affect so
much the change in Reserves
- The amount of payment of Insurance often deviates considerably from the
initial estimate

17
 ITEMS TO BE CONSIDERED:
-APPEARANCE AND PAYMENT
From Accident Occurrence Date (A) to Appearance Date (E) and Payment Date (S)

Objective

A E S
Car Interpersonal

A E S
Worker Compensation

A E S
Product Liability Responsibility

A E S
18
OTHER CONSIDERATIONS
Elements affecting Loss Reserves

Internal and Management Elements

- Reinsurance Scheme
- Insurance Payment Practice
- Growth Potential
- Sufficiency of Case Loss Reserve
- Commodity Composition
- Underwriting Assessment
- Contract Change
- Properties of the Portfolio

19
OTHER CONSIDERATIONS
Elements affecting Loss Reserves

External and Environmental Factors

- Society
- Regulation
- Justice System
- Seasonality
- Residual Market
- Inflation
- Economy

20
BASIC CALCULATION METHODS OF LOSS RESERVES

Loss Development Loss development
FundamentalFundamental
Techniques Techniques

21
BASIC CALCULATION METHOD OF LOSS RESERVES
- DEFINITION

Loss Development
Economic Change related to the claim until the final payment of the insurance from the
occurrence of the claim

Triangle
In order to estimate the future payment patterns, measure the changes in the cumulative
claims

Loss Development Factor

Ratio of the estimates of the successive loses of the certain limited claim group
(ex. accident year,)

22
BASIC CALCULATION METHOD OF LOSS RESERVES
CUMULATIVE PAYMENT INSURANCE LOSS TRIANGLE

Losses are categorized according to the year in which the

accident occurred

Losses are totaled at the end of the year

The losses paid up to the present time are displayed on the

diagonal line

This data is created to analyze the past payment patterns

23
BASIC CALCULATION METHOD OF LOSS RESERVES
CUMULATIVE INSURANCE LOSS TRIANGLE

Framework of Calculation

24
BASIC CALCULATION METHOD OF LOSS RESERVES
CUMULATIVE PAYMENT INSURANCE CLAIM LOSS TRIANGLE

The purpose of this table is to estimate the Total Amount of Final Payment Insurance

25
BASIC CALCULATION METHOD OF LOSS RESERVES
PAYMENT INSURANCE CLAIM LOSS DEVELOPMENT FACTOR

From the end of the accident year (12 months) to the end of the next accident year (24
months), the payment of the insurance in the 1997 accident year has increased by
79%. For the following year (from 24 months to 36 months), the payment insurance
increased by 24%. After that, it is the same. The Loss development Factor is also
expressed as follows:
- Age-to-Age factors
- Link Ratios
26
BASIC CALCULATION METHOD OF LOSS RESERVES
PAYMENT INSURANCE LOSS DEVELOPMENT FACTOR

27
BASIC CALCULATION METHOD OF LOSS RESERVES
APPLICATION OF PAYMENT INSURANCE LDM

28
BASIC CALCULATION METHOD OF LOSS RESERVES
PROJECTION AND LOSS RESERVE BY PAYMENT INSURANCE

Loss Reserve Estimate @ December 31, 2011 = 32.241 million

29
BASIC CALCULATION METHOD OF LOSS RESERVES
PAYMENT INSURANCE MONEY THINGS TO CONSIDER REGARDING LDM

EXAMPLE
Things to be considered
Those in the old accident year include
Has there been any change that disrupted motorcycle accident payment than now.
the continuity between the one of the old Currently it does not handle motorcycle
accident year and the current one? insurance

Are there any things that should not be We began to expand our business in new
taken into consideration or what need to be areas
adjusted?
At the end of December of a certain year,
there was great ice rain damage suffered.
Due to the delay in reporting, we had an
impact not normally seen in the
Development Factor of the next fiscal year

30
BASIC CALCULATION METHOD OF LOSS RESERVES
CUMULATIVE INCURRED LOSS TRIANGLE

31
BASIC CALCULATION METHOD OF LOSS RESERVES
SELECTED INCURRED LDF

32
BASIC CALCULATION METHOD OF LOSS RESERVES
INCURRED LDM PROJECTION, RESERVES FOR LOSS RESERVE

33
PART

34
CONTENTS TO COVER IN PART 2
Comparison of Paid and Incurred Results
Validation
Ultimate Loss Ratios
Frequency/Severity
Pure Premium
Sensitivity Analysis of Current Fiscal Year
Claim Severity
Frequency/Severity Projection

Importance of Tail Factor

35
 COMPARISON OF LDM FORECASTS

Accident Paid Incurred Average =
LDM LDM Selected
Paid Method Incurred MethodAverage
1996 11,244 11,250 11,247
1997 12,985 12,738 12,862
1998 15,215 14,471 14,843
1999 17,588 16,308 16,948
2000 19,109 17,539 18,324
2001 21,435 20,119 20,777

97,576 92,425 95,001

25,000
20,000
15,000
10,000
Paid
5,000 Method
- Incurred
Method
1996 1997 1998 1999 2000 2001 Average

36
COMPARISON OF LOSS DEVELOPMENT METHODS

PLDM: No Change in Payment Pattern

Premise of Assumption ILDM: No change of the Adequacy of Case Loss Reserves

PLDM: Certainty Data; No estimate for data to be used

Advantage ILDM: We use all available data

PLDM: There is a possibility of generating excessive and

fluctuating wide Development Factors. It can also take a long time
Difficult Point until the final forecast

ILDM: We use the case loss reserve as the result of the estimate for
predicting final reserves

37
IMPORTANT ASSUMPTION
POTENTIAL PROBLEMS

Assumption Case of Problem

The pattern of claims payment does not change Improvement on adequacy of Case Loss Reserve
Introduction of new process of Case Loss Reserve
No change of normal reserve fund reserve method

Change of data processing method

No change in complaint processing process Improvement on method of recording claim payment
Increase of maximum amount payment for Insurance
Payment
The upper limit of insurance money does not affect Loss Changing the upper limit
development

38
IMPORTANT ASSUMPTION
POTENTIAL PROBLEMS
Assumption Problem example
No change in product portfolio Rapid Inflation
Increase in Litigation
Decrease in Defense Expenses
We do not assume cyclic Loss development
Change Reinsurance Cover
Increase Long-Tailed Events
It does not assume abnormal value data Partial Refinement of New Collateral Type or Collateral
Type

Insurance Payment, Reserves Funded under the Insurance

underwriting Cycle

When a catastrophe or abnormal Insurance Payment is

included in the actual value of Insurance Payment
Payment for abnormal claims, Delay in reporting

39
COMPARISON OF QUOTATION RESULTS
Estimated Loss Reserves Based on:
Accident Paid Incurred Average =
Year LDM LDM Selected
Paid Method Incurred MethodAverage
1996 736 742 739
1997 1,449 1,202 1,326
1998 2,757 2,013 2,385
1999 4,889 3,609 4,249
2000 7,937 6,367 7,152
2001 14,473 13,157 13,815

Total 32,241 27,090 29,666

Components of Selected Reserve

@ 12/31/01
16,000

14,000

12,000

10,000
IBNR
8,000
Case
6,000

4,000

2,000

Accident
40 Year
 FORMULAS LEADING IBNR RESERVES
It is easy to calculate IBNR if prediction of Final Insurance is
obtained

IBNR Reserve Estimated Final Insurance Payment Insurance Ordinary Loss Re

Estimated Final Insurance Report Insurance Money
Accrued Insurance Payment Ordinary Loss Reserve

41
OTHER FUNDING METHODS

Method to explain in a later part

1. Expected Loss Technique
2. Bornhuetter-Ferguson Method
3. Severity/Frequency Method

There are many other methods

Note) The Development Method also applies to the number of complaints

42
DEPLOYMENT METHOD FOR NUMBER OF COMPLAINTS

Cumulative Number of Claims Reported * Final

Accident Development Stage in Months # of
Year 12 24 36 48 60 72 Claims

1996 1,428 2,772 2,850 2,866 2,870 2,888 ???

1997 1,710 3,032 3,086 3,094 3,110 ???
1998 1,358 2,780 2,990 3,000 ???
1999 1,510 2,588 2,656 ???
2000 1,488 2,604 ???
2001 1,604 ???

43
NUMBER OF REPORTS :
SELECTED LDF
Evaluation Interval in Months
Accident 72 to
Year 12-24 24-36 36-48 48-60 60-72 Ultimate
1996 1.941 1.028 1.006 1.001 1.006 ???
1997 1.773 1.018 1.003 1.005
1998 2.047 1.076 1.003
1999 1.714 1.026
2000 1.750
2001
Simple Average - All Years
1.845 1.037 1.004 1.003 1.006

Simple Average - Latest 3 Years

1.837 1.040 1.004 XXX XXX

Simple Average - Excluding High & Low

1.821 1.027 1.003 XXX XXX

Selected Development Factors

1.821 1.037 1.006 1.004 1.006 1.000

Selected Development Factors to Ultimate

1.919 1.054 1.016 1.010 1.006 1.000

44
NUMBER OF REPORTS :
PREDICTION

Actual Estimated Estimated

Claims Development Ultimate Unreported
Accident Reported Factors to Claims Claims
Year @ 12/31/01 Ultimate [(2) x (3)] (4) - (2)
(1) (2) (3) (4) (5)

1996 2,888 1.000 2,888 0

1997 3,110 1.006 3,129 19
1998 3,000 1.010 3,030 30
1999 2,656 1.016 2,699 43
2000 2,604 1.054 2,744 140
2001 1,604 1.919 3,078 1,474

Total 15,862 17,568 1,706

45
RELEVANCE
The final Insurance should be validated by relevant indicators.

Premium
Loss Ratio

Exposure - or Number of Contracts

Frequency pure premium
Number of Complaints
Loss ratio

Assumptions and Methodology should be recorded as sentences, Sensitivity

Analyzes etc. should be done.

46
 VALIDATION:
FINAL LOSS RATIO
Est. Ultimate Losses ($000) Indicated Loss Ratio
Accident Earned Using: Using:
Year Premium PLDM ILDM Selected PLDM ILDM Selected

1996 18,168 11,244 11,250 11,247 0.619 0.619 0.619

1997 21,995 12,985 12,738 12,862 0.590 0.579 0.585
1998 24,173 15,215 14,471 14,843 0.629 0.599 0.614
1999 25,534 17,588 16,308 16,948 0.689 0.639 0.664
2000 31,341 19,109 17,539 18,324 0.610 0.560 0.585
2001 38,469 21,435 20,119 20,777 0.557 0.523 0.540

Total 159,680 97,576 92,425 95,001 0.611 0.579 0.595

47
 VALIDATION:
FINAL LOSS RATIO

Ultimate Loss Ratio

0.700

0.600
Paid LDM

Incurred LDM

0.500 Selected
1996 1997 1998 1999 2000 2001

Accident Year

48
 VALIDATION:

Ultimate Claim
Accident Claim Earned Frequency
Year Count Exposures* (2) / (3)
(1) (2) (3) (4)

1996 2,888 102 28.314

1997 3,129 98 31.929
1998 3,030 103 29.417
1999 2,699 105 25.705
2000 2,744 109 25.174
2001 3,078 118 26.085

49
 VALIDATION:
SEVERITY/FREQUENCY

Ultimate Est. Ultimate Losses ($000) Indicated Severity

Accident Claim Using: Using:
Year Count PLDM ILDM Selected PLDM ILDM Selected
(1) (2) (3) (4) (5) (6) (7) (8)

1996 2,888 11,244 11,250 11,247 3,893 3,895 3,894

1997 3,129 12,985 12,738 12,862 4,150 4,071 4,111
1998 3,030 15,215 14,471 14,843 5,021 4,776 4,899
1999 2,699 17,588 16,308 16,948 6,516 6,042 6,279
2000 2,744 19,109 17,539 18,324 6,964 6,392 6,678
2001 3,078 21,435 20,119 20,777 6,964 6,536 6,750

50
 VALIDATION:
PURE PREMIUM
Values at 12 Months
Earned Pure
Accident Exposures Reported Premium Percentage
Year ('000's) Losses [(2)/(1)] Change

1996 102 9,337 91.5

1997 98 10,540 107.6 17%
1998 103 11,875 115.3 7%
1999 105 13,343 127.1 10%
2000 109 14,469 132.7 4%
2001 118 16,561 140.3 6%

Estimated Ultimate Values

Earned Pure
Accident Exposures Incurred Premium Percentage
Year ('000's) LDM [(2)/(1)] Change

1996 102 11,250 110.3

1997 98 12,738 130.0 18%
1998 103 14,471 140.5 8%
1999 105 16,308 155.3 11%
2000 109 17,539 160.9 4%
2001 118 20,119 170.5 6%

51
 VALIDATION:
PURE PREMIUM

Pure Premium Trends

200.0
Ultimate Pure

150.0
Premium

100.0

50.0
-
1996 1997 1998 1999 2000 2001

At 12 M onths Additional to Ultimate

52
SENSITIVITY ANALYSIS:
FISCAL YEAR ANALYSIS

The Reason for improving results is ...:

Increase in Premium Rate
Reduction of Claim Frequency
Reduction of Claim Damage Rate
Better results may appear in the following cases:
Claims are handled more slowly and are paid more slowly.
In the case of the satisfaction of Case Loss Reserve is lowered
In the case of the product composition is different

53
SENSITIVITY ANALYSIS:
RATE RELATIONSHIP

Review Past Relationships

Loss
Ratio of Reported Insurance to Payment Insurance
Number of Complaints
Payment Completion Rate
Ratio of No Claim
Loss and Number of Complaints
loss ratio or Average Amount

54
SENSITIVITY ANALYSIS ::
SENSITIVITY ANALYSIS-EXAMPLE:
Ratio of Paid Losses to Reported Losses
Accident Development Stage in Months
Year 12 24 36 48 60 72

1996 0.405 0.615 0.735 0.822 0.889 0.934

1997 0.400 0.618 0.745 0.838 0.907
1998 0.413 0.641 0.772 0.864
1999 0.428 0.661 0.790
2000 0.421 0.666
2001 0.420

Average Reported Loss

Accident Development Stage in Months
Year 12 24 36 48 60 72

1996 6,539 3,913 3,892 3,905 3,915 3,895

1997 6,164 4,025 4,067 4,101 4,092
1998 8,744 4,976 4,762 4,804
1999 8,836 6,005 6,049
2000 9,724 6,442
2001 10,325
55
 SENSITIVITY ANALYSIS :
PERCENTAGE LEVEL SATISFACTORY
Accident Earned Earned Average Change from
Year Premium Exposures Premium Prior Year

1996 18,168 102 178.1

1997 21,995 98 224.4 26%
1998 24,173 103 234.7 5%
1999 25,534 105 243.2 4%
2000 31,341 109 287.5 18%
2001 38,469 118 326.0 13%

The Increase in Average Premium mainly consists of:

Changes in Product Composition
Increase Insurance Premium Rate
If the change in the average premium over the last two years is due to an increase in the rate, the improvement
in the loss ratio can mainly be explained by that.

If the change is mainly due to a change in the composition of the product, the improvement of the loss ratio
may be real, or it may not be the case. It is likely that further investigation is necessary to understand what the
change was and whether that different product type 56
will change the characteristics of Loss development.
SENSITIVITY ANALYSIS :
CLAIM LOSS RATIO
Est. Ultimate Severity Est. Change in Severity
Accident Using: Using:
Year PLDM ILDM Selected PLDM ILDM Selected
(1) (3) (4) (5) (6) (7) (8)

1996 3,893 3,895 3,894

1997 4,150 4,071 4,111 7% 5% 6%
1998 5,021 4,776 4,899 21% 17% 19%
1999 6,516 6,042 6,279 30% 27% 28%
2000 6,964 6,392 6,678 7% 6% 6%
2001 6,964 6,536 6,750 0% 2% 1%

It is a typical example that there is no consistent pattern in the damage rate except that it is
increasing every year in general terms. It is expected to increase due to inflation

57
SENSITIVITY ANALYSIS :
CLAIM END RATE
Accident Number of Closed Claims by Development Age Ultimate
Year 12 24 36 48 60 72 Claims

1996 826 2,131 2,559 2,706 2,795 2,845 2,888

1997 782 2,308 2,738 2,957 3,049 3,129
1998 780 2,146 2,665 2,832 3,030
1999 917 1,980 2,368 2,699
2000 911 1,978 2,744
2001 1,106 3,078

Accident Percentage Closed to Est. Ultimate Number

Year 12 24 36 48 60 72

1996 29% 74% 89% 94% 97% 99%

1997 25% 74% 88% 95% 97%
1998 26% 71% 88% 93%
1999 34% 73% 88%
2000 33% 72%
2001 36%

Example: 29% = 826 / 2,888

Over the past few years, complaints have been completed earlier. This suggests that Insurance
Payments are paid earlier and that the Paid Loss Development Factor - is probably fairly high. This
indicates that one of the important Assumpions of PLDM is collapsing.

58
 SENSITIVITY ANALYSIS:
CASE LOSS RESERVE SUFFICIENCY
Accident Case Reserves ($000)
Year 12 24 36 48 60 72

1996 5,557 4,176 2,936 1,987 1,245 742

1997 6,328 4,664 3,200 2,051 1,189
1998 6,974 4,968 3,251 1,955
1999 7,635 5,274 3,367
2000 8,376 5,604
2001 9,599

Accident Number of Open Claims

Year 12 24 36 48 60 72

1996 602 641 291 160 75 43

1997 928 724 348 137 61
1998 578 634 325 168
1999 593 608 288
2000 577 626
2001 498

Accident Average Case Reserve

Year 12 24 36 48 60 72

1996 9,231 6,515 10,089 12,419 16,600 17,256

1997 6,819 6,442 9,195 14,971 19,492
1998 12,066 7,836 10,003 11,637
1999 12,875 8,674 11,691
2000 14,516 8,952
2001 19,275

59
SENSITIVITY ANALYSIS :
CASE LOSS RESERVE SUFFICIENCY
Generally increasing trend figures are expected:
1. Since smaller claims are paid more quickly, the number increases as you go to the
side of the line
2. Because of inflation, the number increases as you go down the row

It is important to understand the basic funding policies and procedures of case loss
reserves and to be able to interpret trends indicated by data. Many changes related to case
loss reserve procedures can be monitored by communicating with departments
responsible for Insurance Payments.

Changes on the adequacy of case loss reserves will affect ILDM. For example, if the
adequacy of the case loss reserves is somewhat lower in the current accident year,
considerable progress will be seen in terms of the LDM factor than the levels seen in the
past. Using past Loss development factors under these circumstances will lower future
estimates and as a result will be an estimate of insufficient reserves.
60
SENSITIVITY ANALYSIS :
CASE LOSS RESERVE SUFFICIENCY
Average Case Reserve
@ 12 Months
25000
20000
15000
10000
5000
0
1996 1997 1998 1999 2000 2001
Actual Fitted-exp

According to the approximate curve obtained from the average case loss
reserve data every 12 months, the development factor of 19% is shown
annually. This value largely deviates from the industry average price range of
passenger injury insurance from 8% to 10%.

61
FREQUENCY AND LOSS RATIO PREDICTION METHOD
Estimated Ultimate Frequency Estimated Ultimate Severity
Accident Linear Exponential Linear Exponential
Year Actual Fit Fit Actual Fit Fit

1996 28.3 30.6 30.6 3,894 3,740 3,625

1997 31.9 29.4 29.3 4,111 4,347 4,399
1998 29.4 28.1 28.0 4,899 5,051 5,172
1999 25.7 26.9 26.8 6,279 5,870 5,946
2000 25.2 25.6 25.6 6,678 6,821 6,719
2001 ?? 24.4 24.4 ?? 7,927 7,493

R-squared 0.491 0.509 0.954 0.946

Selected for 2001 24.4 7,493

The actual value for the current fiscal year is obtained by approximating the
actual value in the most recent past five years by a straight line or a curve. The
estimated value of the current year is used as an estimate of the final Frequency
and Loss Ratio.

R-aquared represents how well the fit function fits the data.
62
FREQUENCY LOSS RATIO PREDICTION METHOD
Accident Earned Ultimate Ultimate Losses Reserves
Year Exposures Frequency Severity Estimated @ 12/31/01

1996 102 28.3 3,894 11,247 739

1997 98 31.9 4,111 12,862 1,326
1998 103 29.4 4,899 14,843 2,385
1999 105 25.7 6,279 16,948 4,249
2000 109 25.2 6,678 18,324 7,152
2001 118 24.4 7,493 21,618 14,656

Total 635 95,842 30,507

A Comparison of Loss Ratio Methods for 2001

Paid Loss Development Method 55.7%

Incurred Loss Development Method 52.3%

Frequency and Severity Method 56.2%

63
 SELECTION OF TAIL FACTOR
How much does selection of Tail factor influence?

Effect on Estimates Given a 2% Increase in Reported Losses Tail Factor

Reported Estimated Revised Unpaid

Accident Losses Selected LDF's Ultimate Earned Loss Losses
Year @ 12/31/01 LDF Age to Ult. Losses Premium Ratio @ 12/31/01

1996 11,250 1.020 1.020 11,475 18,168 63% 967

1997 12,725 1.001 1.021 12,992 21,995 59% 1,456
1998 14,413 1.003 1.024 14,759 24,173 61% 2,301
1999 16,066 1.011 1.035 16,628 25,534 65% 3,929
2000 16,776 1.030 1.066 17,883 31,341 57% 6,711
2001 16,561 1.162 1.239 20,519 38,469 53% 13,557

Total 87,791 94,256 159,680 59% 28,921

Estimated Unpaid Losses Based on Original ILDM 27,090

(Without the 2% Tail Factor Increase)

Increase in Estimated Unpaid Losses Due to Increased Tail Factor 7%

64
SELECTION OF TAIL FACTOR
Final Loss will increase by 2% or $ 1.8 million
The Reserve will also increase by $ 1.8 million. However, if you
increase the Tail Factor by 2%, the total Reserve increase by
7%.

Regarding to IBNR Reserves, it shows a higher rate of increase for

2% increase in Tail Factor

65
PART

66
CONTENTS TO COVER IN PART 3

1. Expected Loss Ratio Method

2. Defense Expenses and Expenses for Cost Control

(AllocatedLosAdjustment Expenses)

1. Assessment Other expenses(AO)

(Unallocated Loss Adjustment Expenses)

67
EXPECTED LOSS RATIO METHOD

Expected Damage Rate (ELR)

The ratio of the final loss to the anticipated Already-Insured
Premium

Data and Sources to use in anticipation:

Pricing Assumption
Past Data such as schedule P
Industry Data

68
 EXPECTED LOSS RATIO METHOD
Example of ELR using pricing assumpion
Commission 20%

Tax 5%

General expenses 15%

Profit (2%)

Total 38%

Rate and damage investigation expenses amount to be paid 62%

69
EXPECTED LOSS RATIO METHOD
Example of ELR from Schedule P

EZ INSURANCE COMPANY AUTO LIABILITY

Schedule P - Part 1B

Private Passenger Auto Liability/Medical

Years in Loss and Loss Expense Percentage

Which (Incurred/Premiums Earned)
Premiums
Were
Earned
and Losses Direct
Were and
Incurred Assumed Ceded Net

1. Prior XXXX XXXX XXXX

2. 1992 73.1% 73.8% 72.4%
3. 1993 66.6% 65.9% 67.3%
4. 1994 70.3% 68.9% 71.7%
5. 1995 69.0% 70.6% 67.4%
6. 1996 74.1% 75.0% 73.2%
7. 1997 80.2% 83.3% 77.1%
8. 1998 60.5% 59.1% 61.9%
9. 1999 62.6% 61.3% 63.9%
10. 2000 66.7% 68.0% 65.4%
11. 2001 67.0% 68.3% 65.7%

12. Totals XXXX XXXX XXXX

3 year average 65.0%

5 year average 66.8%

70
EXPECTED LOSS RATIO METHOD

Loss reserve estimate based on ELR

Previous insurance premium x ELR = expected final insurance claim

Final Loss - Payment Insurance = Total Loss Reserve

Total reserve for reserves - case reserves = IBNR reserves

71
EXPECTED LOSS RATIO METHOD
Loss Reserve Estimate based on ELR - example
Previous Premium = $100,000
Expected Loss Rate = 0.65
Payment Insurance = $10,000
Case Loss Reserve = $13,000

Loss Reserve = ($100,000 x 0.65) - $10,000

= $65,000 - $10,000
= $55,000

IBNR Reserve = $55,000 - $13,000

= $42,000

72
EXPECTED LOSS RATIO METHOD

Estimate based on Expected Loss Ratio

Use it in the following situations:
New Item
When there was a sudden change in the product line
Long Tail Products and accident year statistics are still shallow

(Final Loss) <(Payment Insurance)

If the calculation result of loss reserve may be negative in this case

73
EXPECTED LOSS RATIO METHOD

Loss Reserve based on Expected Loss Ratio and Loss Reserve based on previously
reported reports
(Previous Premium x ELR) x (IBNR Factor) = (IBNR Reserve)
Here, IBNR actor = (1.000 - 1.000 / LDF *)
Loss Reserve based on previously reported reports + IBNR Reserves = Final insurance
Case Loss Reserves + IBNR reserves = Total Reserve for Loss Reserves

74
EXPECTED LOSSReserves
RATIO METHOD
Based on ELR and Reported Incurred

EZ INSURANCE COMPANY AUTO LIABILITY

(In Thousands)

Expected Cumulative
Accident Earned Loss Expected IBNR Incurred Ultimate
Year Premium Ratio Losses Factor IBNR Losses Losses
(1) (2) (3) (4) (5) (6) (7) (8)
(2) x (3) slide 10 (4) x (5) slide 10 (6) + (7)

1995 17,153 0.60 10,292 0.000 0 10,292 10,292

1996 18,168 0.60 10,901 0.001 11 11,250 11,261
1997 21,995 0.60 13,197 0.002 26 12,725 12,751
1998 24,173 0.60 14,504 0.006 87 14,413 14,500
1999 25,534 0.60 15,320 0.017 260 16,066 16,326
2000 31,341 0.60 18,805 0.046 865 16,776 17,641
2001 38,469 0.60 23,081 0.180 4,155 16,561 20,716

Total 176,833 106,100 5,404 98,083 103,487

75
EXPECTED LOSS RATIO METHOD
Reserves Based on ELR and Reported Incurred

EZ INSURANCE COMPANY AUTO LIABILITY

CUMULATIVE INCURRED LOSSES

(In Thousands)

Accident ---------------- ---------------- DEVELOPMENT STAGE IN MONTHS -------------------------------

----------------
Year 12 24 36 48 60 72 84

1995 8,382 9,781 10,110 10,219 10,268 10,280 10,292

1996 9,337 10,847 11,092 11,192 11,235 11,250
1997 10,540 12,205 12,551 12,690 12,725
1998 11,875 13,832 14,238 14,413
1999 13,343 15,542 16,066
2000 14,469 16,776
2001 16,561

Accident ---------------- ---------------- INCURRED LOSS DEVELOPMENT FACTORS ------------

----------------
Year 12-24 24-36 36-48 48-60 60-72 72-84 84-Ult

1995 1.167 1.034 1.011 1.005 1.001 1.001

1996 1.162 1.023 1.009 1.004 1.001
1997 1.158 1.028 1.011 1.003
1998 1.165 1.029 1.012
1999 1.165 1.034
2000 1.159
2001

ALL YEARS AVERAGE 1.163 1.030 1.011 1.004 1.001 1.001

SELECTED LDFs 1.163 1.030 1.011 1.004 1.001 1.001 1.000

CUMULATIVE LDFs 1.219 1.048 1.017 1.006 1.002 1.001 1.000

1.000
IBNR FACTOR = (1.000 - ------) = % OF EXPECTED LOSSES WHICH ARE UNREPORTED
LDF

IBNR FACTOR 0.180 0.046 76 0.017 0.006 0.002 0.001 0.000

EXPECTED LOSS RATIO METHOD
Comparison of Reserve Methodologies
Expected Losses

Expected $6,000 $10,000

Rptd IBNR

Example : Reported Incurred Losses are Twice as High as Expected

ELR $12,000 $4,000

Rptd IBNR

Bornuetter- $12,000 $10,000

Ferguson Rptd IBNR

Incurred $12,000 $20,000

Development Rptd IBNR

Example : Reported Incurred Losses are Half of Expected

ELR $3,000 $13,000

Rptd IBNR

Bornuetter- $3,000 $10,000

Ferguson Rptd IBNR

Incurrred $3,000 $5,000

Development Rptd IBNR

77
 BORNHUETTER-FERGUSON METHOD
EXCEPT OF PROPERTY INSURANCE APPLICATION EXAMPLE
How many Hideki Matsui will beat Homerun under the following
conditions?

In 40 games, 20 home runs

1 season 160 games
Three pieces of information are required to do the BF method
Expected Final Value
Cumulative Loss Development Factor
Amount occurring up to the present time

78
 BORNHUETTER-FERGUSON METHOD
EXCEPT OF PROPERTY INSURANCE APPLICATION EXAMPLE
Three necessary information in this example:
Hideki Matsui was predicted how many homeruns would beat before the season?
Expected Final Value = 40
To predict season totals from the current statistical data, we currently digit 1/4 games so we can quadruple
the current figures.

Cumulative Loss development factor = 4.000

At the moment we have 20 home runs

Amount occurring so far = 20

79
BORNHUETTER-FERGUSON METHOD
EXCEPT OF PROPERTY INSURANCE APPLICATION EXAMPLE
B-F B-F Projection: Final Value = (Expected Figure * IBNR Factor - Factor) + (Amount occurring up to the
present time)

Factor = 1.000 - (1.000 / LDF) = 0.75 (in other words, 75% games remain)
Final Value = (40 * .75) + 20 = 50
According to the B - F Method, Matsui Hideki will hit 50 homers in one year

0-40 Games 41-80 Games 80-120 Games 121-160 Games 20 Homeruns 10 Home runs
10 Home runs 10 Home runs

80
BORNHUETTER-FERGUSON METHOD
EXCEPT OF PROPERTY INSURANCE APPLICATION EXAMPLE

Comparison between BF Method and other two methods

ILDF Method
Final Expectation Value = Generation amount up to the present time point * Cumulative LDF
= 20 * 4.000 = 80 home run

0-40 Game 41-80 Game 81-120 Game 121-160 Game

20 home runs 20 home runs 20 home runs 20 home runs

Expected loss ratio method

Final Expected Value = Expected Value = 40 Home Runs

0-40 Match 41-80 Game 81-120 Game 121-160 Match

10 home runs 10 home runs 10 home runs 10 home runs

81
BORNHUETTER-FERGUSON METHOD
EXCEPT OF PROPERTY INSURANCE APPLICATION EXAMPLE

Premise Process Problem Case

That pricing does not conform to a

The Premium must accurately reflect the certain methodology
exposure
The loss ratio is not stable in accident
The Expected Loss Ratio must be yearly statistics
estimable
Damage reporting, reserving of reserve,
payment of Insurance Payment etc. Introduction of new complaint handling
must be carried out by certain system
procedures.

82
 EXPECTED LOSS RATIO METHOD ELR)
ELR AND LOSS RESERVE BASED ON PREVIOUSLY
REPORTED REPORTS(BORNHUETTER-FERGUSON )

Advantage Disadvantage

Point of contact between Loss Development Assume that the already Reported Unpaid
and Expected Loss Ratio and the Already Reported are irrelevant
It can be avoid overestimation due to Depends on the accuracy of the Expected
previously generated things that are not Loss Ratio and the reporting pattern,
normally considered
The relevance to the already occurred loss
Applicable to new products that are out of decreases
business line
It depends on the accuracy of Previous
Available without internal loss history premium
Easy to use
83
LOSS INVESTMENT COST (LAE) RESERVE VALUATION
METHOD

Damage Research Expenses are categorized and reported and recorded in

the following two categories
Costs of Defense and Payment Control
Survey Other Cost
Prior to January 1, 1998, whether or not expenses can be assigned to
claims depends on how the costs are recorded and reported in the annual
report as determinants
Alteration of definitions has little effect
DCC is almost same as ALAE
AO is almost the same as ULAE

84
LOSS RESEACH EXPENSE (LAE) RESERVE VALUATION
METHOD
Defense and Payment control costs (DCC)

Costs corresponding to the following items inside and outside the company:
Attorney's Fee
Court Costs
Medical Cost Reduction

Assessment and other expenses (AO)

Cost corresponding to the following items. However, it is not limited to these:

Expenses for agents that are acting on adjusters and payment
Lawyer fee incurred to fix payment amount
Costs involved in litigation between insurance company and contractor
Appraiser, private investigator cost, salary etc.

85
DCC RESERVE CALCULATION METHOD

1. PAID DCC DEVELOPMENT

2. CUMULATIVE PAID DCC TO CUMULATIVE PAID LOSSES

86
DCC RESERVE CALCULATION METHOD
Cumulative Paid DCC
($ in thousands)

EZ INSURANCE COMPANY AUTO LIABILITY

Accident --------------- --------------- DEVELOPMENT STAGE IN MONTHS ------------

---------------
Year 12 24 36 48 60 72 84

1995 71 166 286 416 527 611 677

1996 83 189 313 458 584 672
1997 93 213 361 523 657
1998 103 226 394 581
1999 108 245 437
2000 128 280
2001 132

Accident --------------- --------------- PAID DCC DEVELOPMENT FACTORS ------------

---------------
Year 12-24 24-36 36-48 48-60 60-72 72-84 84-Ult

1995 2.338 1.723 1.455 1.267 1.159 1.108

1996 2.277 1.656 1.463 1.275 1.151
1997 2.290 1.695 1.449 1.256
1998 2.194 1.743 1.475
1999 2.269 1.784
2000 2.188
2001

Average 2.259 1.720 1.461 1.266 1.155 1.108

4 point average 2.235 1.720 1.461

Avg. excl. high/low 2.258 1.720 1.459

Time wght. average 2.239 1.734 1.463 1.264 1.154

Vol. wght. average 2.251 1.724 1.461 1.266 1.155 1.108

SELECTED LDFs 2.251 1.724 1.461 1.266 1.155 1.108 1.108

CUMULATIVE LDFs 10.175 4.520 2.622 1.795 1.418 1.228 1.108

87
DCC RESERVE CALCULATION METHOD
DCC Reserves Based on Paid DCC Development

EZ INSURANCE COMPANY AUTO LIABILITY

($ in Thousands)

Accident DCC Paid Selected Estimated Unpaid

Year to Date Factor Ultimate DCC
(1) (2) (3) (4) (5)
slide 22 slide 22 (2) x (3) (4) - (2)

1995 677 1.108 750 73

1996 672 1.228 825 153
1997 657 1.418 932 275
1998 581 1.795 1,043 462
1999 437 2.622 1,146 709
2000 280 4.520 1,266 986
2001 132 10.175 1,343 1,211

Total 3,436 7,304 3,868

88
DCC RESERVE CALCULATION METHOD

DCC Reserves Based on Paid DCC Development

Advantage Disadvantage
It is similar to Paid Loss, it is Ignoring the relationship with
simple and easy to use Ross

It fits well with the old accident It depends greatly on the initial
year payment amount which is very

89
DCC RESERVE CALCULATION METHOD
Cumulative Paid DCC to Cumulative Paid Losses
($ In Thousands)

EZ INSURANCE COMPANY AUTO LIABILITY

Accident ---------------- ---------------- CUMULATIVE PAID DCC ------------------------------- ----------------

Year 12 24 36 48 60 72 84

1995 71 166 286 416 527 611 677

1996 83 189 313 458 584 672
1997 93 213 361 523 657
1998 103 226 394 581
1999 108 245 437
2000 128 280
2001 132

Accident ---------------- ---------------- CUMULATIVE PAID LOSSES -------- ---------------- ----------------

Year 12 24 36 48 60 72 84

1995 3,361 5,991 7,341 8,259 8,916 9,408 9,759

1996 3,780 6,671 8,156 9,205 9,990 10,508
1997 4,212 7,541 9,351 10,639 11,536
1998 4,901 8,864 10,987 12,458
1999 5,708 10,268 12,699
2000 6,093 11,172
2001 6,962
90
DCC RESERVE CALCULATION METHOD
Cumulative Paid DCC to Cumulative Paid Losses

EZ INSURANCE COMPANY AUTO LIABILITY

Accident ---------------- CUMULATIVE PAID DCC TO CUMULATIVE PAID LOSSES ---------------

----------------
Year 12 24 36 48 60 72 84

1995 0.021 0.028 0.039 0.050 0.059 0.065 0.069

1996 0.022 0.028 0.038 0.050 0.058 0.064
1997 0.022 0.028 0.039 0.049 0.057
1998 0.021 0.025 0.036 0.047
1999 0.019 0.024 0.034
2000 0.021 0.025
2001 0.019

91
DCC RESERVE CALCULATION METHOD
Cumulative Paid DCC to Cumulative Paid Losses

EZ INSURANCE COMPANY AUTO LIABILITY

Accident ---------------- ---------------- PAID TO PAID DEVELOPMENT FACTORS --------------

----------------
Year 12-24 24-36 36-48 48-60 60-72 72-84 84-Ult

1995 1.312 1.406 1.293 1.173 1.099 1.068

1996 1.290 1.355 1.297 1.175 1.094
1997 1.279 1.367 1.273 1.159
1998 1.213 1.406 1.301
1999 1.261 1.442
2000 1.193
2001

Average 1.258 1.395 1.291 1.169 1.097 1.068

4 point avg. 1.237 1.393 1.291

Avg. excl. high/low 1.261 1.393 1.295

Time wght. Average 1.240 1.403 1.291 1.167 1.096

Vol. wght. Average 1.258 1.393 1.291 1.169 1.096 1.068

SELECTED LDFs 1.237 1.393 1.291 1.169 1.096 1.068 1.068

CUMULATIVE LDFs 3.252 2.629 1.887 1.462 1.251 1.141 1.068

92
DCC RESERVE CALCULATION METHOD
DCC Reserves Based on Cumulative Paid DCC to Cumulative Paid Loss Development

EZ INSURANCE COMPANY AUTO LIABILITY

($ In Thousands)

Developed Paid Indicated

Accident Ratio Devel. Paid/Paid Ultimate Ultimate DCC DCC
Year to Date Factor Ratio Losses DCC to Date Reserves
(1) (2) (3) (4) (5) (6) (7) (8)
slide 26 slide 27 (2) x (3) slide 11 (4) x (5) slide 25 (6) - (7)

1995 0.069 1.068 0.074 10,292 762 677 85

1996 0.064 1.141 0.073 11,261 822 672 150
1997 0.057 1.251 0.071 12,751 905 657 248
1998 0.047 1.462 0.069 14,500 1,001 581 420
1999 0.034 1.887 0.064 16,326 1,045 437 608
2000 0.025 2.629 0.066 17,641 1,164 280 884
2001 0.019 3.252 0.062 20,716 1,284 132 1,152

Total 103,487 6,983 3,436 3,547

93
DCC RESERVE CALCULATION METHOD

Advantage Disadvantage
When estimating DCC, there is a
possibility of overestimating or
It also takes into account the
underestimating the estimate of the
relationship between DCC and loss
loss considered
A simple way
It is more complicated than Paid
Provide a means to monitor the DCC Development
relationship between DCC and loss
The initial ratio of DCC and loss is
very wide, but it largely depends on
this
Payment of an important DCC is
often related to complaints where
contracts are terminated without
Insurance Payment
94
Statistical data loses consistency
AO RESERVE CALCULATION METHOD
Assessment and other expenses

Cost corresponding to the following items. However, it is not limited to these

Expenses for agents that are acting on adjusters and payment

Attorneys' fees incurred to settle the payment, including expenses for litigation
between insurance company and contractor

95
AO RESERVE CALCULATION METHOD

50/50Rule
50% of ULAE is paid when the claim amount is uncertain and the
remaining 50% is paid when the claim amount is determined

96
AO RESERVE CALCULATION METHOD

Average of the ratio of Paid AO and payed loss over the past three years
Percentage multiplied by 50% applies to case loss reserves
100% applies to IBNR Reserve
IBNR in a broad sense needs to be divided into Case Loss Reserves and
intrinsic IBNR Development

97
AO RESERVE CALCULATION METHOD

Things to consider when choosing the ratio of Payed AO to Payed loss

It may not be possible to generate an appropriate factor on a 3 year
average:
Payment of AO is not necessarily linked with payment of loss for the
year completely
The selection requires subjective judgment based on the following
matters:
A certain increase trend, the trend factor if there is a declining
trend
Changes in procedure of cost distribution etc
98
AO RESERVE CALCULATION METHOD
Example of "50/50" Rule

EZ Insurance Co. - Auto Liability

($ In Thousands)

Calendar Paid Paid Ratio

Year AO Losses (2) / (3)

(1) (2) (3) (4)

1999 1,038 14,107 0.074

2000 1,244 15,906 0.078

2001 1,459 17,709 0.082

Total 3,741 47,722 0.078

99
AO RESERVE CALCULATION METHOD
Example of "50/50" Rule

Ratio of Paid AO to Paid Losses 0.078

50% of Ratio 0.039

Known Case Loss Reserves 22,989

IBNR Reserve 5,296

AO Reserve

= (0.039 x 22,989) + (0.078 x 5,296)

= 897 + 413

= 1,310
100
AO RESERVE CALCULATION METHOD

Prerequisites for applying 50/50 rules

The age of the claim does not affect the ratio of AO to loss
AO and loss are paid at the same ratio
These assumptions should always be verified when the 50/50
rule is applied.

101