Why risk margin is a part of liability instead of own fund?

[Duc Thinh Vu]

Hi,

I would like to ask why risk margin is a part of liability instead of own fund when it is said that "Risk margin is the compensation required by the insurer for the fact that they have to raise and hold the capital" ?

It is like in the liability side of the balance sheet, there are 2 things: things that the insurer owe and things that belongs to the insurer. As for me, by thinking that way, the risk margin should belong to the equity right ?

 

[Mateusz]

The purpose of risk margin is to bring the liabilities calculated using best estimate assumptions for non-hedgeable risks up to a market value. In SII, your first prescribed step is to calculate the liabilities using best estimate assumptions. But this is not considered market consistent, because theoretically there'd be a 50/50 chance of making a loss by a third party if a portfolio under valuation were to be transferred. Risk margin exists to address this issue.

Then there is a question of how to calculate it. The way I think about it, what methodology we end up adopting does not change the general principle presented above (ie we need the RM because under SII we're required to report the liabilities in a market-consistent way). But in the end we need to place some value on it, and for SII the cost of capital approach is the only option. It does have a nice interpretation because holding capital is what insurers do. The sentence you quote shows the perspective of a third party taking on a block of liabilities - it would indeed need to hold capital for it (that would be part of the SCR), and a compensation for being open to accepting those liabilities can be expressed nicely in terms of the return required on that capital.

 

[Duc Thinh Vu]

The purpose of risk margin is to bring the liabilities calculated using best estimate assumptions for non-hedgeable risks up to a market value. In SII, your first prescribed step is to calculate the liabilities using best estimate assumptions. But this is not considered market consistent, because theoretically there'd be a 50/50 chance of making a loss by a third party if a portfolio under valuation were to be transferred. Risk margin exists to address this issue.

Then there is a question of how to calculate it. The way I think about it, what methodology we end up adopting does not change the general principle presented above (ie we need the RM because under SII we're required to report the liabilities in a market-consistent way). But in the end we need to place some value on it, and for SII the cost of capital approach is the only option. It does have a nice interpretation because holding capital is what insurers do. The sentence you quote shows the perspective of a third party taking on a block of liabilities - it would indeed need to hold capital for it (that would be part of the SCR), and a compensation for being open to accepting those liabilities can be expressed nicely in terms of the return required on that capital.

Hi, thank you very much for your explanation. It helps A LOT.
I still have 2 questions to understand more your answer:
1) What do you mean by "making a loss by a third party if a portfolio under valuation were to be transferred" ? By "third party" here, do you mean the "policyholder" ?

2) By your second paragraph, I understand the RM as the "compensation for holding capital" in the view of a third party who wants to buy the insurance portfolios (i.e. they buy the liabilities and the seller give them the corresponding assets that back these liabilities). However, I think that, if I were in the shoes of the "seller" (i.e. the one who currently owns the portfolio), I myself also have to hold my own capital (in the SCR), so I should be compensated by the "buyer" (because, say, if you buy it, you have to compensate me for the cost I have to bear to keep the portfolio). So, the compensation (i.e. the Risk Margin) should be considered as "something i will receive", so I think it should not be truly a liability, right ?

What do you think about that ? Maybe the thing is that I have no idea about what is going on if a third party comes and buy an insurance portfolio (like what is the process, how the assets/ liabilities are transferred and what is the price of the insurance portfolio).

Sincerely thank you for your kindness!

 

[Mateusz]

Hi, thank you very much for your explanation. It helps A LOT.
I still have 2 questions to understand more your answer:
1) What do you mean by "making a loss by a third party if a portfolio under valuation were to be transferred" ? By "third party" here, do you mean the "policyholder" ?

Hi, by a "third party" I mean a potential entity assuming liabilities of another insurer in a transaction.

Say we valued some liabilities at +1000 on a best estimate basis. If this is set as the transfer price, it means the insurer giving up the liabilities needs to transfer 1000 worth of assets to the entity that takes over the obligations.
If that 1000 earns the risk free rates assumed in the valuation and actual experience (e.g. lapses, mortality etc) unfolds exactly as in the best estimate basis, there will be no gain or loss to the third party. But best estimate means a central scenario - mean of the distribution of possible outcomes - so the odds of experience going worse than assumed (e.g. higher claims need to be paid due to worse mortality, less premiums received due to higher lapses etc.) and making a loss are pretty high. Arguably too high for anyone to accept such a valuation (with no additional margins) as a basis for transfer price.

 

[Duc Thinh Vu]

Hi, by a "third party" I mean a potential entity assuming liabilities of another insurer in a transaction.

Say we valued some liabilities at +1000 on a best estimate basis. If this is set as the transfer price, it means the insurer giving up the liabilities needs to transfer 1000 worth of assets to the entity that takes over the obligations.
If that 1000 earns the risk free rates assumed in the valuation and actual experience (e.g. lapses, mortality etc) unfolds exactly as in the best estimate basis, there will be no gain or loss to the third party. But best estimate means a central scenario - mean of the distribution of possible outcomes - so the odds of experience going worse than assumed (e.g. higher claims need to be paid due to worse mortality, less premiums received due to higher lapses etc.) and making a loss are pretty high. Arguably too high for anyone to accept such a valuation (with no additional margins) as a basis for transfer price.

Ah, it is much clearer for me now. So basically, if we use the Cost of capital method to calculate the RM, we should take the viewpoint of the "potential buyer" of the liabilities (i.e. "the compensation for the BUYER as the buyer has to mobilize his capital") right ?

 

[Mateusz]

Ah, it is much clearer for me now. So basically, if we use the Cost of capital method to calculate the RM, we should take the viewpoint of the "potential buyer" of the liabilities (i.e. "the compensation for the BUYER as the buyer has to mobilize his capital") right ?

Hi, this is how I tend to think about it, yes, but let's wait for an ActEd tutor to confirm this understanding. In practice the calculation is carried out based on the projected SCR of the undertaking holding the liabilities. There is a paper by EIOPA which goes in more detail into the underlying assumptions of this approach - it is assumed, for example, that the total portfolio is transferred to a 'reference undertaking' which does not have any obligations of its own, and that following the transfer it holds the assets equal to the sum of the technical provisions (net of reinsurance) and the SCR. The assets selected are assumed to minimise the market risk SCR so that it can effectively be ignored. These assumptions can be viewed as somehow artificial, but they help understand, I think, why the SCR of the original undertaking can be used in the calculation.

 

[Lindsay Smitherman]

Yes - confirmed! Bear in mind that the capital projected in the RM calculation is a subset of the full SCR, comprising only the SCR components that relate to the non-hedgeable risks. [Effectively, the BEL is already 'market-consistent' in terms of market risk through the use of a risk-neutral valuation approach, ie using risk-free rates for investment returns and discount rates.]

 

[Duc Thinh Vu]

Hi, this is how I tend to think about it, yes, but let's wait for an ActEd tutor to confirm this understanding. In practice the calculation is carried out based on the projected SCR of the undertaking holding the liabilities. There is a paper by EIOPA which goes in more detail into the underlying assumptions of this approach - it is assumed, for example, that the total portfolio is transferred to a 'reference undertaking' which does not have any obligations of its own, and that following the transfer it holds the assets equal to the sum of the technical provisions (net of reinsurance) and the SCR. The assets selected are assumed to minimise the market risk SCR so that it can effectively be ignored. These assumptions can be viewed as somehow artificial, but they help understand, I think, why the SCR of the original undertaking can be used in the calculation.

Thank you very much for your kindness, thank you Lindsay for the confirmation. My last question is that could you please give me the name of the paper of EIOPA that you mentioned in your answer ? Thank you.

 

[Mateusz]

Hi,

1. assumptions about the reference undertaking in EIOPA's rulebook:
https://www.eiopa.europa.eu/rulebook/solvency-ii/article-6376_en

2. old document from 2009 which goes into much more detail.
https://register.eiopa.europa.eu/CE.../CEIOPS-L2-Final-Advice-on-TP-Risk-Margin.pdf

 

[VishalKumar]

Yes - confirmed! Bear in mind that the capital projected in the RM calculation is a subset of the full SCR, comprising only the SCR components that relate to the non-hedgeable risks. [Effectively, the BEL is already 'market-consistent' in terms of market risk through the use of a risk-neutral valuation approach, ie using risk-free rates for investment returns and discount rates.]

Thanks for the above discussions!
Wanted to know that in BEL calculations we don’t take investment income so what investment returns are you referring to.

 

[Lindsay Smitherman]

Hi - for the non-unit part of the BEL for a unit-linked product, for example, we need an assumed investment return in order to roll forward the unit fund to determine the expected future charges (and potentially also the cost of any guaranteed benefits).

 

[VishalKumar]

Hi - for the non-unit part of the BEL for a unit-linked product, for example, we need an assumed investment return in order to roll forward the unit fund to determine the expected future charges.

Thanks Lindsay!
That means for calculating non unit BEL components which are dependent on fund like mortality charge,Fund management charge…we project the future unit fund growth rate with risk free yield and then again take their present values at risk free rate.
But what about unit fund if we are growing and discounting at the same risk free rate that it is better to use its current fund value only for BEL.

 

[Lindsay Smitherman]

Yes - that's right: the 'unit reserve' part of the BEL just = unit fund.

 

[Duc Thinh Vu]

Yes - confirmed! Bear in mind that the capital projected in the RM calculation is a subset of the full SCR, comprising only the SCR components that relate to the non-hedgeable risks. [Effectively, the BEL is already 'market-consistent' in terms of market risk through the use of a risk-neutral valuation approach, ie using risk-free rates for investment returns and discount rates.]

Hi Lindsay,
As stated by Mateusz in his first comment, the Risk margin is calculated to bring the liabilities up to market value, and i think it is reasonable. Then you said that the BEL is already market consistent.

I imply from your statement that because BEL is already market consistent, then RM is not necessary, which seems to me a wrong thought.

Could you please help me to clarify this ? Many thanks!

 

[Lindsay Smitherman]

Hi - have another careful look at exactly what I wrote:
'Effectively, the BEL is already 'market-consistent' in terms of market risk through the use of a risk-neutral valuation approach ...'

The risk margin makes the technical provisions market-consistent in terms of the other risks, ie the non-hedgeable risks.