CHAPTER-3

RATEMAKING
 DATING BACK TO HISTORY :
• Lloyd’s of London based early hull rates in part upn the design and
protection of each specific ship, and the classification assigned to each
vessel was written down or manual for the use by the individual
underwriters.
• 18th century dwelling fire insurance rates in the U.S were based upon
ROOF TYPE and BASIC CONSTRUCTION.
• Early rate manuals – provide general guidance to underwriters in setting
the specific rates, rather than actual rates to be charged
• They contained many of the elements associated with present –day
property and liability rate manuals including regognition of differing loss
costs between classifications , expense provision and provision for adverse
deviation and profit
• NO MATTER HOW REFINED THE CLASSIFICATION AND RATING
PROCESS MAY BECOME , MANUAL RATES ARE STILL ESTIMATE OF
AVERAGE COSTS BASED UPON A COMBINATION OF STATISTICAL
 BASIC TERMINOLOGY
• Exposure : Basic rating unit underlying an insurance premium is
called Exposure.
• WRITTEN EXPOSURE : Those units of exposures on policies written
during the period in question.
• EARNED EXPOSURE: The exposure unit actually exposed to loss
during the period.
• IN-FORCE EXPOSURE: The exposure units exposed to loss at a given
point in time.
The specific exposure unit used for a given type of insurance depends
upon several factors including :
1. Reasonableness : The exposure unit should be a reasonable
measure of the exposure to loss. For example : A 1999 Rolls Royce
and a 1989 cherolet might each represent a car year exposure – the
selected measure should directly relate to loss potential to the
2. EASE OF DETERMINATION : The most reasonable and
responsive exposure definition is of no use if it cannot be
accurately determined.
3. REPONSIVENES TO CHANGE : An exposure unit that reflects
changes in the exposures to loss is preferable to one which
does not.
The exposure unit for workers compensation insurance, which
provides benefits which are keyed to average wage levels , is
PAYROLL. This is obviously preferable to number of employees ,
for example , as payroll will change with the prevailing wage
levels.
4. HISTORICAL PRACTICE : Where a significant body of historical
exposure data is available , any change in the exposure base
could render the prior history unusable.
 Since , ratemaking generally depends – review of past statistical
• CLAIM : Demand for payment by an insured or by an allegedly injured
third party under the terms and conditions of an insurance contract.
• CLAIMANT : The individual making the claim is claimant and there can
be multiple claimants within a single claim.
• ACCIDENT DATE : The date of occurrence which gave rise to the claim
• REPORT DATE : The date the insurer receives notice of the claim.
• FREQUENCY : The number of claims is directly related to the number
of exposure.
Fk = Frequency per k exposure units Fk =
K = Scale factor Kc/E
C = Claim count
E = Exposure units
 LOSSES AND LOSS ADJUSTMENT
• LOSSES – Amount paidEXPENSES
or payable to claimants under the terms of
insurance policies are referred to as LOSSES.
• PAID LOSSES – Are those losses for a particular period that have actually
been paid to climants.
• CASE RESERVE – Representing the estimated amount of that payment.
• ACCIDENT YEAR CASE – LOSSES - The sum of all paid losses and case
reserves for a specific accident year at a specific point in time is known as
the Accident year case incurred losses
• ULTIMATE INCURRED LOSSES – INCURRED LOSS + IBNR
• LOSS ADJUSTMENT EXPENSES : Expenses associated with the
settlement of claims, as distinguished from marketing , investment or
general administrative operations.
• ALLOCATED LOSS ADJUSTMENT EXPENSES: Which can be directly
related to specific claim.
• UNALLOCATED LOSS ADJUSTMENT EXPENSES : Which cannot be
related specific claim.
PAID LOSSES
REPORTED
LOSS
OUTSTANDING
LOSSES
INCURRED
RESERVE FOR LOSSES
CASE
RESERVE
DEVELOPEME
NT UNREPORTED
LOSSES
UNREPORTED
CLAIM
RESERVE
• SEVERSITY: Average loss per claim.
SEVERITY = LOSS / CLAIM COUNT
S = L/C

• PURE PREMIUM : The average loss per unit of exposure.

PURE PREMIUM = LOSS/ EXPOSURE
UNIT

PURE PREMIUM = FREQUENCY *

SEVERITY

• PROFIT AND CONTINGENCIES PROVISION : Reflecting the

fact that profits if any will be based upon actual results and not
expectations or projections. R= Rate per unit of exposures
R = P + F PF = = Pure Premium
Fixed expense per exposure
1-V-Q V = Variable expense factor
Q = Profit and contingencies
factor
• NUMERICAL EXAMPLE :
Loss and loss adjustment expense pure premium : $ 75
Fixed expense per exposure : $ 12.5
Variable expense factor : 17.5%
Profit and contingies factor : 5%
Rate = $75+$12.5 = $112.9
1- 0.175-0.05
• LOSS RATIO : Loss divided by premium
• GOAL OF MANUAL RATEMAKING PROCESS - is to be determined
rates that will, when applied to the exposure underlying the risks
being written , provide sufficient funds to pay expected losses and
expenses; maintain an adequate margin for adverse deviation ; and
produce a reasonable return on (any) funds provided by investors.
• Rates shall not be inadequate , excessive or unfairly discriminatory
between risk of like kind and equality.
STRUCTURE OF THE RATING PLAN
1. Age of insured
2. Gender of insured
3. Marital status of insured
4. Prior driving record of the insured
5. Annual mileage driven
6. Primary use of vehicle
7. Make and model of vehicle
8. Age of Vehicle
9. Garaging location of vehicle
• The structure of various elements involved in the manual rating
of a specific risk in known as RATING PLAN.
• Various specific elements are often referred as
CLASSIFICATION , SUB CLASSIFICATION OR RATING
FACTORS.
• SKIMMING THE CREAM : Exclusion of not processing the
positive characteristic because of high cost.
• ADVERSE SELECTION : if other companies are reflecting the
negative factor in their rating plans , the result will be a tendency
towards insuring risks possessing the negative characteristic ,a
situation known as adverse selection.
• Risk characteristics underlying a manual rating plan can be
identified as :
I. Those generally impacting frequency
II. Those impacting severity
 FREQUENCY SEVERITY

Prior driving record (Individual with recent Some vehicles tend to be more susceptible
automobile accident and traffic violations have , damage in collisions than do the other vehicles.
as a class, higher frequencies of future claims. Repair costs for a Rolls Royce costs more than
do those for a Chevrolet.

Individual driving high – powered sports cars as a The presence or absence of sprinkler system
class , higher frequencies than those driving will impact severity as will the value of the
family sedans. building and contents being insured.

Annual mileage driven

In commercial fire insurance , restaurants

generally have a higher frequency than do
clothing stores.

Workers compensation statistics detail higher

frequencies for manufacturing employees than for
 THE RATEMAKING PROCESS
• There are 2 basic approaches to addressing the problem of
manual ratemaking : the PURE PREMIUM METHOD and THE
LOSS RATIO METHOD.
• PURE PREMIUM METHOD
• The pure premium method develop indicated rates.
• Those rates that are expected to provide for the expected losses
and expenses and provide the expected R =profit
RATE PER– UNIT
based upon
OF EXPOSURE
P = PURE PREMIUM
formula F = FIXED EXPENSE PER EXPOSURE
• R = P + F VQ == VARIABLE EXPENSE FACTOR
PROFIT AND CONTINGENCIES
1-V-Q FACTOR
 LOSS RATIO METHOD
• The loss ratio method develops indicated rate changes rather
than indicated rate.
• Indicated rates are determined by application of an adjustment
factor , the ratio of the exposure loss ratio to a target loss ratio ,
to current rates.
• In mathematical terms the loss ratio mthod
R = Indicated rate works as follows :
R0 = Current rate
R = AR0 W=L
A = Ajustment factor = W/T
W = Experience loss ratio
T = 1-V-Q ER0 T = Target loss ratio
V= premium – related expense factor
Q = profit and contingencies factor
1+G
G = ratio of non – premium – related expenses to
losses
E = EXPERIENCE PERIOD EARNED EXPOSURE
R0 = Current rate
R = L(1+G)
 Relationship Between PURE
PREMIUM and LOSS RATIO METHOD
• The pure premium and loss ratio method will produce
identical rates when applied to identical data and using
consistent assumption.
PURE PREMIUM METHOD LOSS RATIO METHOD
BASED ON EXPOSURE BASED ON PREMIUM
DOES NOT REQUIRE EXITING REQUIRE EXISITNG RATES
RATES
DOES NOT USE ON-LEVEL USES ON-LEVEL PREMIUM
PREMIUM
PRODUCES INDICATED RATES PRODUCES INDICATED RATES
• Pure premium method requires well – defined , responsive exposures.
• Loss ratio method cannot be used for a new line
• Pure premium method is preferable where on-level is difficult calculate.
1. Pure premium method requires well – defined , responsive exposures.
• The pure premium method based on losses per unit exposure.
• Where the exposure unit is not available or is not reasonably consistent between
risks , as in the case of commercial fire insurance , the pure premium method
cannot be used.
2. Loss ratio method cannot be used for a new line
• Because the loss ratio method produces indicated rate changes , its use requires
an established rate and premium history
• Where manual rates are required for a new line of business and assuming there
are relevant loss statistics available , the pure premium method be used.
• Of course, if no statistical data are available , then neither method ca n be used.
3. Pure premium method is preferable where on-level premium is difficult
to calculate
• In some instances , such as commercial lines where individual risk rating
adjustments are made to individual policies , it is difficult to determine the on-
level earned premium required for the loss method.
 NEED FOR COMMON BASIS
• Whichever ratemaking method is selected , the actuary needs to make
certain that the experience losses are on a basis consistent with the
exposures and premiums being used.
SELECTION OF EXPERIENCE PERIOD
• Determination of loss experience period to be used in the manual
ratemaking process involves a combination of statistical and
judgement elements.
DIFFERENCE IN COVERAGE
• Wherever possible , major coverages within a line of insurance are
generally treated separately.
• For example : Liability experience under homeowners policies is often
reviewed separately from the property experience.
• Auto collision data is usually analyzed separately by deductible.
• Professional liability – written on claims made basis is different from
 TREATMENT OF INCREASED LIMITS
• Liability coverage rate manuals generally provide rates for a basic
limit of liability along with increased limits factors to be applied to
those base rates where higher limits are desired.
ON LEVEL PREMIUM – ADJUSTMENT FOR PRIOR RATE CHANGES
• The experience period extends over several years there have typically
been changes in manual rate levels between the beginning of the
experience period and date as of which rates are being reviewed.
• If the actuary is using the loss ratio method in the development of the
indicated rate level changes , the earned premium underlying the loss
ratio calculations must be on a current rate level basis
• Doing this manually – time consuming – software is available , the
resulting on-level premiums will be quite accurate. This method is
referred to as the Extension of exposure technique
• When extension of exposures cannot be used, an alternative is called
the PARALLELOGRAM METHOD , is available.
• This method adjusts calendar year earned premiums to current rate
level based upon single geometric relationships and an undelying
assumption that exposure is uniformly distributed over time.

1.00
0
1.125

01/1996 07/1996 01/1997 07/1997

01/1998

• The average 1997 relative earned rate level = [(0.125)*(1.00) +

(0.875)*(1.125)] = 1.1094
• Current relative average level = 1.2375
• Calender year earned premium = 1.2375/1.1094 = 1.1155
In cases where material changes in
exposure level have occurred over the
period , or where there is non-uniform
pattern to written exposures, the
parallelogram method may not produce
a reasonable approximation of on-level
earned premium.
 THE LOSS DEVELOPMENT
METHOD
• The loss development method is based upn the
assumption that claims move from unreported to
reported and unpaid to paid in a pattern that
sufficiently consistent that past experience can be used
to predict the future development.
DEVELOPMENT
CHANGE IN
EXPOSURE

T ES
LEVEL

A
IO ND
AT
LU
EVA
• SETTLEMENT LAG : The period between loss reporting and loss payment.
• REPORTING LAG : The period between loss occurrence and loss reporting –
which affects both claims and losses.
IDENTIFICATION TRENDS
• Once claims and losses have been projected to an ultimate basis it is
necessary to adjust the data for any underlying trends that are expected to
produce changes in indications between experience period and the period
during which manual rates will be in effect.
• The most obvious trend affecting the ratemaking data is the trend in
severity.
• Monetary inflation – increases in jury awards , and increase in medical
expenses are examples of factors that cause upward trends in loss
severities.
• Frequency is also subject to trend.
• Court decisions may open new for litigation that would increase liability
frequencies
• Legal and social pressures might reduce the incidence of driving under the
• Some exposure bases also exhibit identifiable trends
• Workers Compensation uses payroll as an exposure base and products liability
coverage might be based upon dollars of sales.
• As automobile prices increase , physical damage premiums will reflect the change ,
even though no rate change has been made.
• When using the loss ratio method for ratemaking it is ratemaking it is important
that the effect of such trends on premium be properly reflected.
• Frequency and severity trends are often analyzed separately.

REFLECTION OF TRENDS
• Linear , y = ax+b or
• Exponential , y = beax
• Linear model – analogous to simple interest
• Exponential Model – analogous to compound interest.
• The use of a linear model will produce negative values at some point in the future.
• The use of a linear model over an extended period in such cases is generally
inappropriate since frequency , severity , pure premium and exposure must all
greater than or equal to zero.
 TREND BASED UPON EXTERNAL DATA
• Where sufficient loss or claim experience to produce reliable trend
indications not available , the actuary might supplement or supplant the
available experience with external data.
• Insurance trade associations , statistical bureaus and the U.S
Government produce insurance and general economic data regularly.
• Masterson (1968) provides a good general reference on the subject.
• Lommele and Strugis (1974) provide an interesting example of the
application of economic data to the problem of forecasting workers
compensation insurance result.
TREND AND LOSS DEVELOPMENT – “ THE OVERLAP FALLACY”
• It has occasionally has been suggested that there is a double –counting
of severity trend in the ratemaking process where both loss
development – factors – which reflect severity changes as development
on unpaid claim – and severity trend are applied to losses.
• Cook dealt with this subject in details , and with elegance , in a 1970
paper.
 ILLUSTRATION

$
1 WRITTEN PREMIUM 11,540,000.00
$
2 EARNED PREMIUM 10,832,000.00
INCURRED LOSS AND ALLOCATED $
3 LAE 7,538,000.00
$
4 INCURRED ULAE 484,000.00
$
5 COMMISONS 1,731,000.00
$
6 TAXES , LIECENSE & FEES 260,000.00
$
7 OTHER ACQUISITION EXPENSES 646,000.00
$
8 GENERAL EXPENSES 737,000.00

$
TOTAL LOSS AND EXPENSE 11,396,000.00

T= 1-V-Q/1+G
G 0.064208013 ULAE / ALAE
RATIO OF COMMISION TO
WRITTEN 0.15 COMMISION / WRITTEN PREMIUM
RATE OF TAXES , LIECENSES &
FEES TO WRITTEN 0.022530329 TAXES , LIECENSE & FEES / WRITTEN PREMIUM
RATIO OF OTHER ACQUISITION TO OTHER ACQUISITION EXPENSES / WRITTEN
 PROFIT AND CONTINGENCIES
• The profit and contingencies provision represents the essence of
insurance in that it is designed to reflect the basic elements of
risk and rewards associated with the transaction of insurance
business.
• UNDERWRITING PROFIT : Premium collected exceed the
expenses and losses paid , the insurer makes what is called an
underwriting profit
• UNDERWRITING LOSS : When expense exceed premium
• INVESTMENT PROFIT : Arising out of the investment of funds
between premium collected and payment of expenses and
losses.
 PROFIT PROVISIONS IN MANUAL
RATES
• Until the mid-1960s insurance rates would typically include a
profit contingencies provision of approximately 5% of premium.
• The 5% provision produced sufficient underwriting profits to
support the growth of the industry , and it was not generally
viewed as being excessive.
RISK ELEMENTS
• RISK LOADING : A portion of the profit and contingencies
provision represents a provision for adverse deviation or a RISK
LOADING.
• PARAMETER RISK : Parameter risk is simply the risk associated
with the selection of the parameters underlying the applicable
model of the process.
• PROCESS RISK : It is the risk associated with the projection of
future contingencies that are inherently variable.
 RUIN THEORY APPROACH
• The development of a probabilistic model of the insurance
operation and then , generally through MONTE CARLO
SIMULATION , determining the probability of ruin (insolvency)
over a fixed period of time.
• A maximum acceptable probability of ruin is then determined
and the rate level assumption underlying the model is adjusted
to the minimum rate level producing a ruin probability less than
or equal to the acceptable level.
OVERALL RATE INDICATIONS
• The determination of the overall average indicated rate change
will be made on the basis of the experience losses, expense
provisions , profit and contingencies provisions and , in case of
the loss ratio method , on level premium.
 ILLUSTRATION

EXPERIENCE LOSS AND ALLOCATED - ACCIDENT 231637

1 YEAR (1997-99) 51
318114
2 ON LEVEL PREMIUM -CALENDER YEAR 1997-99 48
EXPERIENCE LOSS AND ALLOCATED RATIO 0.72815
3 [(1)/(2)] 8

4 TARGET LOSS AND ALLOCATED RATION 0.6611

RATE CHANGE INDICATION FOLLOWS DIRECTLY

1.10143
INDICATED OVERALL RATE LEVEL CHANGE [3/4] 4
 CREDIBLITY CONSIDERATION
• It is only necessary to understand that a statisitical indication I 1 has an
associated credibility z , between 0 and 1 , relative to some other indication I 2.
• The resulting CREDIBLTIY – WEIGHTED INDICATION I1,2 is determined by the
formula :
I1,2 = z (I1) + (1-z) (I2)
FOR EXAMPLE :
Overall rate indication of +7.28% is 0.85
Alternative indication , for the source , of +4.5% , the credibility –weighted
indication would be 6.86% , determined as follows :
(0.85)(0.0728)+(0.15)(0.0450) = 0.0686
INDICATED CLASSIFICATION RELATIVITIES
The relationship between that rate for a given classification (or territory) to the
base rate is the classification (or territorial) relativity.
The determination of indicated classification relativities is similar to the process
used in the overall rate level analysis.
 INCREASED LIMITS
• While the level of attention to the development of rates for
increased limit is sometimes less than that given development of
basic limit rates.
• The number of increased limit factors that exceed 2.00 should
serve to focus attention on this important element of manual rate
making.
TRENDING INDIVIDUAL LOSSES
• This method involves the application of severity trend to a body
of individual loss data.
• CLOSED CLAIM data are used in order to avoid the
problems associated with projecting loss development
Resulting
distribution of on
individualAnnual
claims.
severity Trend factor
trended closed
claim is then used
trend is applied to each to determine
• Applicationdetermined
of this methodclosed
require
claim unlimited
appropriate losses as
projection base. increased
factor.
limit
LOSS DEVELOPMENT LAYER
• Loss development patterns by layer.
• This process involves seggregationg case-incurred loss data by
policy limit and loss layer and then tracking the observed loss
development factor in each layer.
• Generally sparsity of data in the upper limit precludes the use of
this method.

FITTED SIZE LOSS DISTRIBUTION

• Related to individual loss trending method.
• A Theoretical size-of-loss distribution is fitted to exisitiing
individual loss data.
• Resulting distribution – used to- examine – effects of severity
trend on various limits and as a basis for the increased limit
factors.
THANK YOU !!!!!!!!!!