Use of Catastrophe Models Flashcards

1
Q

Differences between Catastrophe modelling and traditional rating (5).

A
  • Traditional rating works for high frequency, low severity. CAT modelling is for low frequency, high severity.
  • For CAT, observed losses may not reflect true underlying risk.
  • The return period is large e.g. 250 year return.
  • Uses computing power and Geographical Information Systems.
  • Most models in infancy stage. Increase in sophistication in short time.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

CAT modelling: Approach

A
  • It starts with historical events but applys a longer timescale ( decades or
    century)
    -The model uses past experience and scientific understanding to create possible future events and stochastic event sets ( never happened).
    -The model calculates the effect of events on the insured portfolio.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Basic structure of CAT models (5).

A
  1. Event module- details of stochastic events defined by physical parameters, location and likelihood of occurrence. Based on seismological and meteorological assessment. Uncertainty: incomplete event set, missing extremes, ENIDS etc
  2. Hazard module- determines the hazard of each event at each location. The hazard is the consequence of the event. E.g wind speed ( hurricane) , ground shaking ( earthquake). Based on seismological and meteorological assessment. Uncertainty: errors in hazard model e.g low wind speed
  3. Inventory / exposure module- details of insured’s risk and risk factors e.g location, age, occupancy, occupation. Rely on input data specific to the user of model. Uncertainty: user error in exposure or incomplete
  4. Vulnerability module- degree of loss to a structure resulting from exposure to given hazard -% of TSI. Based on engineering assesment. Uncertainty: under/over estimate losses e.g construction types respond differently, demand surge not modelled
  5. Financial analysis module- Uses database of policy conditions (limits, excess )to translate the total ground up loss to insured losses. Rely on input data specific to user of model. Uncertainty: coverage not correctly modelled e.g excluding flooding
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Perils modelled under CAT

A
  1. Hurricane
    - US, Carribbean, N Europe, Japan, Far East and Latin America.
  2. Earthquake
    - US, Japan, Canada, S Europe, Mexico and Indonesia.
  3. Others- tornado, hail, river flood , wild fire etc.
  4. Non natural/ man-made
    - Terrorism models
    - Deterministic module- maximum expected loss from different terrorist attack.
    - Or stochastic module
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Infectious Disease modules

A
  • recently developed
  • Assumptions-needed on changes in demographics, ease and frequency of travel and improvement on medical care.
  • not commonly used
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Parameters for Hurricanes in event set.

Typhoons and cyclones

A
  • Track (path hurricane follows)
  • Maximum sustained speed
  • Storm radius
  • Forward speed
  • the rate of decay of the wind field
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Parameters for Earthquake within event sets.

A
  • the moment magnitude (measuring energy release)
  • Focal depth - shallow faults are more damaging
  • total area of fault rupture
  • fault type - convergent plate high hazard
How well did you know this?
1
Not at all
2
3
4
5
Perfectly