CP1 - Chapter 28

[Darragh Kelly]

Hi,

I have several questions in relation to chapter 28.

Section 2.1 Scenario Analysis

I was looking at the question 7 from the paper 1 2022 September exam and the description of scenario analysis from page 6 of chapter 28.

I was looking for some clarify/further explanation on the following:

1. Are the scenarios identified before the risk exposures are grouped in broad categories? In q7 of the September 2022 paper it starts with scenario identification then follows with putting the various risk exposures into broad categories, but it in the notes (page 6 ch28) it does it the opposite way around.

2. What exactly is a risk exposure? Is an example like interest rates, inflation rates, survival rates etc?

3. Is a risk factor the same as a risk exposure? What exactly is a risk factor?

4. Can one scenario be sensitive to several different risk exposures ie market risks is the risk exposure and then you might have identified the scenarios: 1. collapse of financial markets and 2. Extremely high inflation which are both sensitive to market risks?

5. Is the total cost of the scenario analysis the sum of the cost of each individual scenario?

Section 2.4 Reserves Stress testing pg 9

On page 9 there is a paragraph which covers business plan failure needs. What in this context does it mean when it says ‘to cover it’s min risk appetite’?

Section 2.5 Stochastic modelling pg 10

On Page 10 the there is a bullet point that describes an approach running different stochastic variable followed by a single deterministic runs, which determine the effect of interactions between variables. I’m not sure I follow how this works?

Section 4.2 Liability Risks pg 16

It is important to stress the need for consistent…

…of the population exposed to risk

For example, a life insurer might need to analyse…

…to ensure correspondence in the exposed to risk analysis.

Just wondering could I get a bit more detail on these paragraphs as find hard to follow what it means thanks.

Section 5.3 Reporting at enterprise level pg19

If two business units are allocated risk exposures that diversify away…

…additional capital will need to be held to cover the unbalanced risks taken on.

Just wondering could I get a bit more detail on these paragraphs as find hard to follow what it means thanks.

Many thanks,

Darragh

 

[James Nunn]

Hi Darragh

Some answers to you Chapter 28 questions are below - hopefully this helps:

Section 2.1 Scenario Analysis

Regarding your questions:

1. It can be done either way.

2. Yes - they can be risk exposures. Risk exposures considered under scenario analysis are often those that are more difficult to assess with a mathematical model though, as covered in this section of the notes.

3. They are not the same thing in this contact - the risk exposure is the real-life risk and the risk factor is how you're covering/parameterising this in a model.

4. Yes - scenarios will generally involve a number of risk exposures.

5. The total cost of the scenario analysis the sum of the cost of all risks included for a particular (ie single) scenario? (This probably makes more sense given my answer to 4 above.)

Section 2.4 Reserves Stress testing pg 9

A company won't just hold capital because it has to due to regulation - it will also want to do this to avoid the risk of failure. The lower it's risk appetite is for failure, the more capital it will want to hold.

Section 2.5 Stochastic modelling pg 10

Under this approach, you'd learn more about - for example - the impact of a 2% tail event occurring (worst outcome with a 2% probability), for each variable by running a model (or models) that models (or model) variables one by one stochastically (one run of the model for each), with the other variables being modelled deterministically. You could then run a deterministic model with all variables as best estimate, then with all parameters adjusted to be in line with the values of parameters that are assessed as being the worst case likely to occur with a 2% probability. (The change in variable sizes between runs would be decreased from those determined earlier to reflect correlations between variables.) The difference between the results for these two fully deterministic runs would then give you a combined impact using an alternative method to stochastically modelling all variables at the same time.

Section 4.2 Liability Risks pg 16

when you calculate actual or expected rates of an occurrence, you'll divide the number of occurrences by some measure of the exposed to risk (the number of things that could result in an occurrence). If these two aren't consistent, the rate you calculate will not be correct. Also, you can only compare actual and expected occurrences if they are the actual and expected numbers of the same type of occurrence, otherwise the comparison will be meaningless as you're not comparing like with like.

Section 5.3 Reporting at enterprise level pg19

For example, an overall assessment of a company's risk may assess that the aggregate of two risks is less than the sum due to a negative correlation between risks. This would mean that total risk exposure could be greater if one risk is less than expected as there would be less of an offset.

For example - annuity business and life insurance business. A company may have volumes of each of these types of business such that mortality risk largely offsets longevity risk. However, if the annuity business decreases or is less than thought, then less of the mortality risk will be offset and total (longevity plus mortality) risk could increase.