Modelling Flashcards

1
Q

What is a model?

A

*small representation of reality
*designed to show NB features of reality

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Model sources

A

*New
*Modify existing model
*external commercial purchase

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

How to decide on model

A

FACEUP
F: flexibility required
A: accuracy required
C: cost
E: expertise in house
U: use of model
P: fit for purpose

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Dynamic model

A

Allows for interaction between various features under different circumstances

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Deterministic model

A

SEER
S: clarity on scenarios tested
E: Easy and quick to design, build and run
E: easy to explain results
R: requires thought as to range of economic scenarios to run

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Stochastic models

A

COUS

C: allow for CORRELATION between variables
O: good at assessing OPTIONS and GUARANTEES
U: better at showing UNCERTAINTY of results
S: test a wide range of SCENARIOS

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Process of modelling

A
  1. Specify purpose of model
  2. Collect, group, modify data
  3. Choose model form
    3s1 - choose suitable density function for each stochastic variable
    3s2 - specify correlations between variables
    3d1 - identify parameters and variables
    3d2 - ascribe parameter values
  4. Construct model based on expected cashflow
  5. Check goodness of fit
    6 . If GoF fails fit new model
    6d1 - Run model using selected values
    6d2 - sensitivity test parameters
    6s1 - run model repeatedly using random generated values of stochastic variables
    6s2 - produce summary of results
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Specify purpose of model

A
  1. Set overall time horizon
    *future period of protections
  2. Set time for output of cashflows
    *frequent - reliable output
    * too frequent- spurious accuracy
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Model point(s)

A

*representative single policy representing a homogeneous group of policies
*capture most important characteristics of the group
* output scaled up to get numbers for whole group
*chosen to reflect expected profile of future business

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Discount rate

A
  1. Allows for
    * required return
    *statistical risk attaching to the cashfows
    -model, parameter and random fluctuation risk
    - assessed analytically (sensitivity analysis, stochastic model, comparison with market data)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Discount rate theory vs practice

A

Theory: diff discount rate for each component of net cashflow
Reality: single rate, simple, reflects average level of risks

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Premiums

A
  1. Model points
  2. Risk discount rate
  3. Marketability of premiums
    * design
    *features
    *sales channel
    * profit criterion
    *size of market
    *decision to market contract
  4. Options and Guarantees
    *stochastic model
  5. Capital requirements
    6.busin3ss objectives
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Good model

A

TERRIFIC DRIP

T: time period between projected cashflows
E: consisted with general economic principles
R: reflect risk profile
R: run must not be long, complex, expensive with hard to interpret results
I: implementable in a range of ways
F: include significant FEATURES
I: input parameters and values must be appropriate
C: communicable and independent verifiable output
D: valid, rigorous and well DOCUMENTED
R: refinement and development
I: interactions must be sensible
P: parsimonious

How well did you know this?
1
Not at all
2
3
4
5
Perfectly