4. Modélisation du capital économique et agrégation des riques Flashcards

(11 cards)

1
Q

Strategies are continuously employed to reduce model risks. Name 3 situations where such strategies are used

A
  1. When choosing a model for the task
  2. When overseeing the usage of the model
  3. When communicating the results of the model
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

What are 4 ways to quantify the potential severity of the failure of a model?

A
  1. Financial significance of the results
  2. Importance of decisions based on the model
  3. Frequency of use
  4. Non-financial impact (ex: reputation)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

What are 4 ways to quantify the potential likelyhood of the failure of a model?

A
  1. Complexity of the model
  2. Expertise of users
  3. Documentation
  4. Sufficiency of testing
  5. Independence of the one who validated the model
  6. Peer review
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Define model risk rating

A

Model risk rating is a way to evaluate the model risk using a certain rating method. It can be uni-directionally or bi-directionally

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

CIA’s educational note proposes a 4-steps process when choosing a new model to use, what are those 4 steps? (The model already exists)

A
  1. Review specifications (are the assumptions appropriate?)
  2. Validate implementation (test and compare model)
  3. Deal with limitations (understand the limitations and act in consequence)
  4. Document model choice
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

If an actuary wants to use a model outside of his area of expertise, how can he make sure that the model will be appropriate?

A
  1. Verify that those who created and validated the model had appropriate knowledge of their field. The actuary might also consider a risk rating of the model.
  2. The actuary would still try to understand the basic of the model like the inputs, outputs, the general validation that was done, the model’s complexity and control framework.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

When preparing the documentation on how to use a model, what are 6 aspects that should be well documented?

A
  1. Instructions to obtain input data
  2. Authorization to set input assumptions
  3. How to run the model
  4. Checks to be applied
  5. Reconciliation from prior runs
  6. Flowchart of the process
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

To use a model, it is important to validate it first. What are the 4 things that should be validated?

A
  1. The data used in the model
  2. The assumptions used in the model
  3. The results from the model
  4. The documentation of the model
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

To validate data and make sure it is both reliable and sufficient, what can the actuary do?

A

Sufficient:
- Data meets model specification

Reliable:
- Reconciliated with other sources
- Summarized and compared to prior periods
- Outliers are investigated
- Missing data is handled

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

How can the actuary review the results of the model?

A

Make sure the results are reasonable, expected and consistent. If results are not one those three, further investigation is needed. Results that are not expected are not necessarilly bad, but we need to understand why we didn’t expect it.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

If the modelling is incidental to the engagement, does the actuary need to disclose the model?

A

No, unless there are limitations that need to be disclosed

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