uum-bwrr3033-risk management--chapter 04 risk measurement

Upload: rodziah-ahmad

Post on 04-Jun-2018

225 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    1/29

    1

    Chapter 4: Risk Measurement

    Frequency and severity of losses

    Probability

    Probability distribution Fault tree

    Pooling arrangement and diversification of

    risk

    BWRR3033_Rodziah(c)

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    2/29

    Learning Objective

    Review the concepts of probability and statistics

    Apply mathematical concepts to understand the

    frequency and severity of losses

    Understand the concepts of expected value andvariance of random variables

    Distinguish between binomial distribution and

    poisson distribution, and which is more appropriatefor different situations

    Show how pooling of independent loss exposures

    reduces risk

    BWRR3033_Rodziah(c) 2

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    3/29

    RISK MEASUREMENT

    BWRR3033_Rodziah(c) 3

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    4/29

    Random Variables & Probability Distribution

    BINOMIAL

    DISTRIBUTION

    POISSONDISTRIBUTION

    PROBABILITY TREE

    DIAGRAM

    RV is a variable which outcome is uncertain. Example: Flipping a coin.

    Each flips result is uncertain.

    Probability Distribution identifies all of the

    possible outcomes, associates a probability with

    each outcome.

    BWRR3033_Rodziah(c) 4

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    5/29

    0

    0.1

    0.2

    0.3

    0.4

    0.5

    0.6

    RM0 RM500 RM1,000 RM5,000 RM10,000

    Probability

    Probability

    BINOMIAL

    DISTRIBUTION

    POISSONDISTRIBUTION

    Possible Outcomes for Damages()

    Probability

    1 RM0 0.50

    2 RM500 0.30

    3

    RM1,000 0.10

    4 RM5,000 0.06

    5 RM10,000 0.04

    The question now

    How we make decision based on these data?

    Use

    Expected Value

    Variance or Standard Deviation

    Skewness

    Correlation

    BWRR3033_Rodziah(c) 5

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    6/29

    Expected Value

    it is where the outcomes tend to occur, on

    average.

    ,

    =

    BWRR3033_Rodziah(c) 6

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    7/29

    Expected Value

    Probability

    Outcome

    B A

    BWRR3033_Rodziah(c) 7

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    8/29

    Variance & Standard Deviation

    =

    Variance

    Standard Deviation

    =

    BWRR3033_Rodziah(c) 8

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    9/29

    Expected Value

    Probability

    Outcome

    B

    A

    BWRR3033_Rodziah(c) 9

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    10/29

    Expected Value & Skew

    Probability

    Outcome

    D

    C

    BWRR3033_Rodziah(c) 10

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    11/29

    MPL : Worst loss in worst scenario

    PML: Worst loss in normal scenario

    Maximum Probable Loss

    Vs

    Probable Maximum Loss

    BWRR3033_Rodziah(c) 11

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    12/29

    Maximum Probable Loss

    (Value At Risk)

    Probability

    Annual Liability Loss RM20m

    area= 0.05

    Area = 0.01

    RM30m

    BWRR3033_Rodziah(c) 12

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    13/29

    Area: 0.05

    Value at Risk

    Probability

    -RM7.5m -RM5m

    Area: 0.01

    0

    Monthly change in value of portfolio

    BWRR3033_Rodziah(c) 13

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    14/29

    When discussing about many types of risk, it is

    important to study the relationship among

    random variables.

    Correlation = 0the random variables are

    not related.

    :: outcome of one random variables will not

    give info about the other random variables.

    :: random variables are said to be

    independent or uncorrelated.

    Correlation

    BWRR3033_Rodziah(c) 14

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    15/29

    REVISION

    INDEPENDENT

    EVENTS

    MUTUALLY

    EXCLUSIVE

    EVENTS

    COLLECTIVE

    EXHAUSTED

    EVENTS

    ALTERNATIVE

    EVENTSJOINT EVENTS

    BWRR3033_Rodziah(c) 15

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    16/29

    Pooling Arrangements and

    Diversification of Risk Basic Idea:

    Replace your loss with the average loss of a group

    Issues:

    What happens to each persons Expected loss

    Standard deviation of loss

    Maximum probable loss

    How do these results change with More participants

    Correlation in losses among the participants increases

    BWRR3033_Rodziah(c) 16

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    17/29

    Uncorrelated Risk Pooling Example

    with 2 People Two people with same distribution (w/o pooling)

    Outcome Probability

    $2,500 0.20

    Loss =$0 0.80

    Assume losses are uncorrelatedExpected value =

    Standard deviation =

    BWRR3033_Rodziah(c) 17

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    18/29

    Uncorrelated Risk Pooling Example

    with 2 People Two people with same distribution (w/o pooling)

    Outcome Probability

    $2,500 0.20

    Loss =$0 0.80

    Assume losses are uncorrelatedExpected value =

    2500 0.2 + 00.8

    $500Standard deviation =

    2500500 0.2 + 0 500 0.8 $1000

    BWRR3033_Rodziah(c) 18

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    19/29

    Uncorrelated Risk Pooling Example

    with 2 People

    Loss

    No

    Loss

    Loss

    Loss

    No

    Loss

    No

    Loss

    0.2

    0.8

    0.2

    0.2

    0.8

    0.8

    Tree Diagram Loss Amount to be shared

    2500 x 2/2 = 2500

    2500 / 2 = 1250

    2500 / 2 = 1250

    0

    BWRR3033_Rodziah(c) 19

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    20/29

    Uncorrelated Risk Pooling Example

    with 2 People Pooling Arrangement changes distribution of

    accident costs for each individualOutcome Probability

    $0

    Cost = 1,250

    2,500

    Expected Cost =

    SD =

    BWRR3033_Rodziah(c) 20

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    21/29

    Uncorrelated Risk Pooling Example

    with 2 People Pooling Arrangement changes distribution of

    accident costs for each individualOutcome Probability

    $0 0.64

    Cost = 1,250 0.32

    2,500 0.04

    Expected Cost = $500

    SD = $707

    BWRR3033_Rodziah(c) 21

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    22/29

    Effect on Expected Loss

    w/o pooling, expected loss = $500

    with pooling, expected loss = $500

    Effect on Standard Deviation

    w/o pooling, standard. deviation = $1,000

    with pooling, standard. deviation = $707

    Risk Pooling Example with 2 People

    BWRR3033_Rodziah(c) 22

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    23/29

    Risk Pooling Example with

    More People

    BWRR3033_Rodziah(c) 23

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    24/29

    The Effect of Risk Pooling Arrangements on Probability

    Distributions for a large number of small business

    Without Pooling

    With Pooling

    P

    robabilityDistributions

    Cost paid by each business20000 40000 60000BWRR3033_Rodziah(c) 24

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    25/29

    Risk Pooling of Uncorrelated Losses

    Main Points: Pooling arrangements

    do not change expected loss

    reduce uncertainty (variance decreases,losses become more predictable, maximum

    probable loss declines)

    distribution of costs becomes moresymmetric (less skewness)

    BWRR3033_Rodziah(c) 25

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    26/29

    Effect of positive correlation on Risk Reduction

    Pro

    babilityDistrib

    utions

    Loss

    Positive Correlated

    Uncorrelated

    BWRR3033_Rodziah(c) 26

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    27/29

    Effect of positive correlation on Risk Reduction

    Standard

    Deviation

    ofAverageLoss

    Number of participants in risk pooling

    Uncorrelated

    Less than perfect correlation

    Perfect Correlation

    BWRR3033_Rodziah(c) 27

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    28/29

    RISK MEASUREMENT (Analyze)DEFINITION

    (What?)

    ESTIMATION

    (How)

    LOSS

    FREQUENCY

    Number of times loss occurs

    during a specific time on record.

    nil/ slight/ moderate/ ultimate

    Very likely/ likely/ unlikely

    LOSS

    SEVERITY

    Size of loss per occurrences. Critical/ important/

    unimportant fatal,/ major/ minor/

    negligible

    BWRR3033_Rodziah(c) 28

  • 8/13/2019 UUM-BWRR3033-Risk Management--CHAPTER 04 Risk Measurement

    29/29

    RISK MEASUREMENT (Evaluate)

    BWRR3033 Rodziah(c) 29