OSFI - Earthquake Flashcards
What are the best practices for earthquake modeling? (1-4)
- DOCUMENT use of the model within risk management program
- explain why a particular model is used versus ALTERNATIVES
- QUALIFIED STAFF needed to run in-house models regularly
- must have KNOWLEDGE of Assumptions, Methods, Limitations of the model
What are the best practices for earthquake modeling? (5-7)
- provide evidence to show that the granularity and quality of the DATA is appropriate
- understand how UNCERTAINTY affects the capital adequacy and the reinsurance requirements
- if PML(model 1) <> PML(model 2), explain the differences, and mention the subsequent model adjustments (where PML = Probable Maximum Loss)
What are the 5 key principles for managing earthquake exposure?
1) Risk management
2) Data management
3) Models
4) PML (Probable Maximum Loss)
5) Financial resources and contingency plan
What is the broad 3-point plan for managing earthquake exposure?
- MEASURE
- MONITOR
- LIMIT EQ EXPOSURE
Briefly describe the key principle of RISK MANAGEMENT for EQ exposure.
EQ exposure Risk Management policies are SUBJECT TO OVERSIGHT by the Board of Directors and IMPLEMENTED by senior management.
Briefly describe the key principle of DATA MANAGEMENT for EQ exposure.
- data required is MORE than for traditional ratemaking
- must address the INTEGRITY, VERIFICATION, LIMITATIONS
Briefly describe the key principle of MODELING for EQ exposure.
One must understand the assumptions, the methods, and the limitations of the EQ models.
Briefly describe the key principle of PML (Probable Maximum Loss) for EQ exposure.
PML = Total Expected Ultimate Cost
- includes considerations for data quality, non-modeled exposure, model uncertainty, and multi-region exposure
Briefly describe the key principle for FINANCIAL RESOURCES & CONTINGENCY PLAN for EQ exposure.
- Financial Resources: quantification of how financial resources cover PML (Probable Maximum Loss)
- Contingency Plan: how to continue efficient business operations after a disaster
Identify uses of EQ models aside from PML (probable Maximum Loss) calculation. (2)
- assist in U/W decisions
- monitor exposure accumulation
What are sound practices for EQ model VERSION? (4)
- use more that 1 model
- ensure timely updates of material changes to the model (within 1 year of the change)
- understand the assumptions, methods, and limitations of the vendor software for PML (Probable Maximum Loss) calculation
- if in-house PML model is used, one should compare the results to alternate models
What are sound practices for EQ model VALIDATION? (3)
- compare modeled losses with actual losses
- compare tail losses with market price for reinsurance
- use global data to supplement limited Canadian EQ data
How might management adjust for low data quality in an EQ PML (Probable Maximum Loss) estimate?
They may add a margin of safety to the PML estimate (though, this is not an excuse to ignore the data quality).
Identify non-modeled exposures when calculating the PML (Probable Maximum Loss). (4)
- exposure growth between the date of data and the relevant exposure period
- consider the adequacy of the ITV (Insurance-to-Value)
- consider GRC (Guaranteed Replacement Cost), may be inadequate due to inflation
- increased seismicity after a large event
Identify 2 examples of model uncertainty.
- uncertainty associated with conversion from location-specific ground motion to actual damage levels
- model assumptions are being continuously updated & refined