SII EEV and MCEV

[gd1991]

Hi,

I was reading through Chapter 20 and think I have confused myself about EEV under SII. Below, I'm ignoring any PVIF which might exist in SII to keep it simple. Do you disagree with any of the following:

For SII MCEV, there is no PVIF. The RM is the shareholder's allowance for risk so the EV is the Surplus + SCR. I thought this was reasonable as the firm could hedge its liabilities at any time with another insurer for a cost of (BEL+RM) and then the SCR and Surplus is immediately available for the shareholders.

For SII non-market consistent EV, there is still no PVIF. EV = Surplus + SCR + RM - Cost of holding SCR and RM = Surplus + Value of SCR and RM releases

(because the RM and SCR are treated like "required capital"?)

I may be wrong and only one of the above is valid?

 

[Lynn Birchall]

Hi,

I was reading through Chapter 20 and think I have confused myself about EEV under SII. Below, I'm ignoring any PVIF which might exist in SII to keep it simple. Do you disagree with any of the following:

For SII MCEV, there is no PVIF. The RM is the shareholder's allowance for risk so the EV is the Surplus + SCR. I thought this was reasonable as the firm could hedge its liabilities at any time with another insurer for a cost of (BEL+RM) and then the SCR and Surplus is immediately available for the shareholders.

For SII non-market consistent EV, there is still no PVIF. EV = Surplus + SCR + RM - Cost of holding SCR and RM = Surplus + Value of SCR and RM releases

(because the RM and SCR are treated like "required capital"?)

I may be wrong and only one of the above is valid?

Hello

Good question - the future of embedded values under Solvency II is pretty uncertain. Some companies have announced their intention to no longer publish an EV (the argument goes along the lines "if Solvency II provides a market-consistent valuation of the business, why is there then a need to publish a different one?"). Some companies intend to carry on with EV, arguing that Solvency II isn't how they would choose to do a market-consistent valuation, and so an EV that doesn't have to comply with contract boundaries, matching adjustments, a risk margin calculated using 6% etc better reflects their value.

So, one general thing I'd say is that there's unlikely to be a unique answer to the question of what an EV should look like under the Solvency II regime!

In terms of your specific suggestions (& carrying on with your assumption of ignoring and PVIF), I think I'd be tempted to define EV as surplus + SCR + RM - (cost of holding SCR + RM) for both types of EV. Although the RM is not available to the shareholders immediately, the expectation is that it will be released over time.

A main difference between a market-consistent and a non-market consistent version of this EV would be in the interpretation of what "cost of holding capital" meant. For a market-consistent EV this might mean frictional costs, eg tax. For a non-market consistent EV this might reflect the cost of deferral, ie the use of a higher discount rate with a risk margin.

Hope these comments help. Best wishes
Lynn

 

[gruhaa]

Hello

Good question - the future of embedded values under Solvency II is pretty uncertain. Some companies have announced their intention to no longer publish an EV (the argument goes along the lines "if Solvency II provides a market-consistent valuation of the business, why is there then a need to publish a different one?"). Some companies intend to carry on with EV, arguing that Solvency II isn't how they would choose to do a market-consistent valuation, and so an EV that doesn't have to comply with contract boundaries, matching adjustments, a risk margin calculated using 6% etc better reflects their value.

So, one general thing I'd say is that there's unlikely to be a unique answer to the question of what an EV should look like under the Solvency II regime!

In terms of your specific suggestions (& carrying on with your assumption of ignoring and PVIF), I think I'd be tempted to define EV as surplus + SCR + RM - (cost of holding SCR + RM) for both types of EV. Although the RM is not available to the shareholders immediately, the expectation is that it will be released over time.

A main difference between a market-consistent and a non-market consistent version of this EV would be in the interpretation of what "cost of holding capital" meant. For a market-consistent EV this might mean frictional costs, eg tax. For a non-market consistent EV this might reflect the cost of deferral, ie the use of a higher discount rate with a risk margin.

Hope these comments help. Best wishes
Lynn

HI Lindsay
I have a one question. Why we calculate cost of capital for holding risk margin which is itself the cost of holding SCR?

 

[Lindsay Smitherman]

Bear in mind that the risk margin (RM) is only the cost of holding part of the SCR.

Let's think first about the purpose of the RM. The BEL allows for market risk through use of risk-free rates for investment returns and discounting. However, it does not allow for other types of risk, hence an adjustment has to be made: the addition of the RM.

There are many different ways in which a market-consistent liability can be adjusted to allow for non-market risks. For example, the non-market assumptions (mortality, expenses etc) can have margins included. EIOPA chose instead to use what we call the 'cost of capital' method, which means determining the cost of holding capital to support those non-market risks. This is just one way by which an approximate allowance can be made for the non-market risks within the liability cashflows.

So, when we are thinking about holding provisions, the total amount that we should hold is the BEL + RM: a market-consistent liability which allows for both market and non-market risks, and which thus represents the amount that another insurer would require in order to take on the liability.

Now let's think about calculating an embedded value. This means determining the value of the business to the shareholder. The RM and SCR will both be locked into the business: the shareholder cannot receive them immediately. Therefore it can be argued that the assets backing the RM and SCR have less than market value worth to the shareholder - particularly if there are frictional costs which will erode their value (eg taxation, investment expenses) during the time that they are locked in. As Lynn says above, for a non-market-consistent EV, the erosion of value comes from the fact that the assets will earn an assumed investment return which is lower than the discount rate (or the shareholders' required rate of return). The deduction made to the market value to allow for the lock-in value erosion is the 'cost of capital' adjustment. It reduces the embedded value (the value of the business to the shareholder).

So the 'cost of capital' in each of these cases represents a slightly different thing. In the Solvency II balance sheet, it (the RM) represents an allowance for non-market risks. In the EV, it represents the erosion of value from having capital locked into the company without being able to distribute it to shareholders immediately.

However, there is a relationship. If a company decided that it would define its EV 'cost of capital' as being the RM, then (ignoring with-profits business, which makes things more complicated):
EV = free surplus + required capital (SCR + RM) - cost of holding capital (RM)
= free surplus + SCR = own funds.
This is as explained in Chapter 19 of the notes, and is consistent with the concept of technical provisions as being the amount that another insurer would require to take on the liabilities.

 

[gruhaa]

Bear in mind that the risk margin (RM) is only the cost of holding part of the SCR.

Let's think first about the purpose of the RM. The BEL allows for market risk through use of risk-free rates for investment returns and discounting. However, it does not allow for other types of risk, hence an adjustment has to be made: the addition of the RM.

There are many different ways in which a market-consistent liability can be adjusted to allow for non-market risks. For example, the non-market assumptions (mortality, expenses etc) can have margins included. EIOPA chose instead to use what we call the 'cost of capital' method, which means determining the cost of holding capital to support those non-market risks. This is just one way by which an approximate allowance can be made for the non-market risks within the liability cashflows.

So, when we are thinking about holding provisions, the total amount that we should hold is the BEL + RM: a market-consistent liability which allows for both market and non-market risks, and which thus represents the amount that another insurer would require in order to take on the liability.

Now let's think about calculating an embedded value. This means determining the value of the business to the shareholder. The RM and SCR will both be locked into the business: the shareholder cannot receive them immediately. Therefore it can be argued that the assets backing the RM and SCR have less than market value worth to the shareholder - particularly if there are frictional costs which will erode their value (eg taxation, investment expenses) during the time that they are locked in. As Lynn says above, for a non-market-consistent EV, the erosion of value comes from the fact that the assets will earn an assumed investment return which is lower than the discount rate (or the shareholders' required rate of return). The deduction made to the market value to allow for the lock-in value erosion is the 'cost of capital' adjustment. It reduces the embedded value (the value of the business to the shareholder).

So the 'cost of capital' in each of these cases represents a slightly different thing. In the Solvency II balance sheet, it (the RM) represents an allowance for non-market risks. In the EV, it represents the erosion of value from having capital locked into the company without being able to distribute it to shareholders immediately.

However, there is a relationship. If a company decided that it would define its EV 'cost of capital' as being the RM, then (ignoring with-profits business, which makes things more complicated):
EV = free surplus + required capital (SCR + RM) - cost of holding capital (RM)
= free surplus + SCR = own funds.
This is as explained in Chapter 19 of the notes, and is consistent with the concept of technical provisions as being the amount that another insurer would require to take on the liabilities.

Hi Lindsay
Thanks a lot for helping me understand the RM. I have found my understanding about Risk Margin under SII, lets denote it as RM(SII), has been quite different. And i just this last round of discussions around it. Below is the plot:
1.As you mentioned, RM(SII) is the prudence for non-hedgable risk calculated as frictional cost of holding SCR capital for these risk using 'Cost of Capital' approach. that means: RM(SII) =Prudendnce for non-hedgable risk(NHR) = frictional cost ot SCR.
My understanding was, RM(SII) is Prudence for NHR + Cost ot holding SCR for NHR risks.
Basis for my such understanding is based on the content of two pages of Chapter 19, as given below:
page 18, section MCEV, content under 'Component of an MCEV' :
These are
1.free surplus(princ 4)
2.required capital(SCR?, if based on SII, as per prinp 5)
2(a) minus "Frictional cost of required capital"
3. ViF:
3(a) PVFP
3(b) minus Time value of option and guarantee
3(c) minus cost of residual non-hedgeable risks-RM(EEV), when using MCEV.
You must see 2(a) and 3(c) are separate things.
Page 20, EV reporting under SII:
Points written as
(i) if there is no PVIF(assuming item 3(c) above is taken as 2(b) and release of RM(SII) is clubbed with release of SCR)
(ii) company considering RM(SII) to be appropriate measure of cost of residual NHR, RM(EEV), and lock in cost of holding regulatory capital(SCR)
The EV equals to SII own funds.

Hope you can see the reason of my confusion. Can you help me on this?

 

[Lindsay Smitherman]

I think you are getting confused because you are seeing something as being equal to the Solvency II RM when it only might be equal to the RM (and, in reality, probably wouldn't be the same - unless the company was deliberately taking short-cuts).

Under Solvency, the RM represents the allowance for non-market risk. The methodology is proscribed, representing the run-off of the SCR for non-hedgeable risks only, multiplied by a fixed 6% parameter.

Under a MCEV, the calculation needs to allow for two things:
- the frictional cost of locking in capital that cannot otherwise be distributed
- an allowance for non-market risk.

The company performing the MCEV calculation has discretion over what approaches it takes to determine each of these - there are no rules to follow.

As an example, the company might decide to base the first of these (cost of required capital) on the frictional lock-in cost of holding the SCR and the RM (if the EV calculation is being done using the Solvency II balance sheet as the starting point) since both of these items represent amounts that cannot be immediately distributed. The company might decide to allow for the frictional lock-in cost of capital by applying a cost of capital parameter of 1% per annum, say, to the run-off of these amounts. Its choice of this parameter is completely at its own discretion, as is its choice of method for allowing for this lock-in cost.

It might then decide to allow for the second of these (non-market risk) by adding a small prudence (or 'risk') margin to each non-market assumption that it uses, eg a slightly higher per policy expense assumption etc. The approach that it takes to non-market risk adjustment is again entirely at its own discretion. It could alternatively, say, take the same calculation approach as used for the Solvency II Risk Margin, but using its own choice of parameter - likely materially lower than the 6% prescribed by EIOPA.

Rather than doing the above two separate calculations, it could possibly be the case that the company decides that, rather than doing all of this extra work, it could just simplify the whole thing and say that its Solvency II Risk Margin is approximately equal to the sum of the above - sufficient to cover non-market risk and cost of lock-in. In which case (and assuming that PVIF = 0) the EV would equal own funds. This is an approximation, but it basically then just means that there is no point doing loads of extra work calculating an EV for disclosure purposes when analysts can get a pretty good idea of shareholder value (and how it changes over time) just by looking at own funds in the Solvency II balance sheet.

Where a company is more likely to do the calculation more carefully would be where PVIF is not zero (eg contract boundaries in place) or for a portfolio transaction scenario where the EV calculation is driving the amount paid, and so needs to be more accurate - particularly where a company believes that the 6% per annum cost of capital prescribed by EIOPA is too high.

Hope that helps. I think you need to think about the Solvency II Risk Margin as being just a proxy calculation that is designed to reflect non-market risk approximately within the supervisory solvency balance sheet, bearing in mind that for Solvency II you have to follow rules to calculate it and use specified parameters. But for MCEV companies can take a wide range of different approaches to allowing for non-market risk and capital lock-in.

 

[gruhaa]

I think you are getting confused because you are seeing something as being equal to the Solvency II RM when it only might be equal to the RM (and, in reality, probably wouldn't be the same - unless the company was deliberately taking short-cuts).

Under Solvency, the RM represents the allowance for non-market risk. The methodology is proscribed, representing the run-off of the SCR for non-hedgeable risks only, multiplied by a fixed 6% parameter.

Under a MCEV, the calculation needs to allow for two things:
- the frictional cost of locking in capital that cannot otherwise be distributed
- an allowance for non-market risk.

The company performing the MCEV calculation has discretion over what approaches it takes to determine each of these - there are no rules to follow.

As an example, the company might decide to base the first of these (cost of required capital) on the frictional lock-in cost of holding the SCR and the RM (if the EV calculation is being done using the Solvency II balance sheet as the starting point) since both of these items represent amounts that cannot be immediately distributed. The company might decide to allow for the frictional lock-in cost of capital by applying a cost of capital parameter of 1% per annum, say, to the run-off of these amounts. Its choice of this parameter is completely at its own discretion, as is its choice of method for allowing for this lock-in cost.

It might then decide to allow for the second of these (non-market risk) by adding a small prudence (or 'risk') margin to each non-market assumption that it uses, eg a slightly higher per policy expense assumption etc. The approach that it takes to non-market risk adjustment is again entirely at its own discretion. It could alternatively, say, take the same calculation approach as used for the Solvency II Risk Margin, but using its own choice of parameter - likely materially lower than the 6% prescribed by EIOPA.

Rather than doing the above two separate calculations, it could possibly be the case that the company decides that, rather than doing all of this extra work, it could just simplify the whole thing and say that its Solvency II Risk Margin is approximately equal to the sum of the above - sufficient to cover non-market risk and cost of lock-in. In which case (and assuming that PVIF = 0) the EV would equal own funds. This is an approximation, but it basically then just means that there is no point doing loads of extra work calculating an EV for disclosure purposes when analysts can get a pretty good idea of shareholder value (and how it changes over time) just by looking at own funds in the Solvency II balance sheet.

Where a company is more likely to do the calculation more carefully would be where PVIF is not zero (eg contract boundaries in place) or for a portfolio transaction scenario where the EV calculation is driving the amount paid, and so needs to be more accurate - particularly where a company believes that the 6% per annum cost of capital prescribed by EIOPA is too high.

Hope that helps. I think you need to think about the Solvency II Risk Margin as being just a proxy calculation that is designed to reflect non-market risk approximately within the supervisory solvency balance sheet, bearing in mind that for Solvency II you have to follow rules to calculate it and use specified parameters. But for MCEV companies can take a wide range of different approaches to allowing for non-market risk and capital lock-in.

yes, lindsay. Your explanation has finally cleared my confusion around Risk Margin. thanks a lot for your time in such a busy period.

 

[Lindsay Smitherman]

Great - I am so glad that this has helped. No problem at all - you're welcome!

 

[Mbotha]

This is an approximation, but it basically then just means that there is no point doing loads of extra work calculating an EV for disclosure purposes when analysts can get a pretty good idea of shareholder value (and how it changes over time) just by looking at own funds in the Solvency II balance sheet.

Hi Lindsay

I've been thinking about this and, even though EV is similar to the SII balance sheet results (particularly if PVIF= 0) under SII regulations, wouldn't the need for EV still depend on the method used to calculate liabilities in the annual accounts?

E.g. If companies are still using the Old UK GAAP accounting principles for insurance contracts, then the liabilities included in the accounts (whether on an IFRS or New UK GAAP basis) would be based on the more prudent SI reserves. As such, the accounting profits wouldn't give a true view of shareholders' interests and we'd need to calculate an EV (using these more prudent SI reserves) to provide a more accurate view. The regulatory balance sheet would need to reflect SII liabilities though but, because this is only disclosed to the regulator, there would be no need to calculate another (different) EV using these SII liabilities - is that right?

So, in the example above, a company ends up having to do two sets of liability calculations and potentially two sets of EV calculations (if the second EV mentioned above is required). It seems like far less work to just align the accounts to reflect SII liabilities (if these can be seen to be more relevant and no less reliable). Would you agree?

 

[Mbotha]

Hi Lindsay

I just want to make sure I understand MCEV correctly...

MCEV = (1) PVIF + (2) RC - (3) cost of RC + (4) Free Surplus

Please can you confirm my understanding of the following:

1. The cost of residual non-hedgeable risk (ie non-market risk) must be allowed for in (1) and this could be done either (a) via margins in the best estimate assumptions used to project the cashflows or (b) by including a risk margin in the projection of the releases in reserves. In the case of (b), the risk margin doesn't have to be calculated using the prescribed 6% under SII.
2. (2) will include the SCR, the RM (if its run-off isn't included in the PVIF) and may also include RC to meet internal objectives. How does the company decide when to include the latter?
3. (2) is just the value of these capital amounts at the valuation date - ie no need to project them and discount the releases because the discount rate and investment return assumptions are the same in MCEV
4. Will the RM in (2) be the SII RM (ie calculated using the 6%)? If not, in which situations will it not be this?
5. (3) requires all the elements of (2) to be projected so that a chosen cost of capital rate (not necessarily 6%) can be applied to those projected capital amounts (and then discounted). This would be the frictional cost of capital

Thanks in advance for your help!!

 

[Lindsay Smitherman]

I've been thinking about this and, even though EV is similar to the SII balance sheet results (particularly if PVIF= 0) under SII regulations, wouldn't the need for EV still depend on the method used to calculate liabilities in the annual accounts?

Companies may start EV calculations from the supervisory balance sheet rather than the accounting balance sheet.

The regulatory balance sheet would need to reflect SII liabilities though but, because this is only disclosed to the regulator, there would be no need to calculate another (different) EV using these SII liabilities - is that right?

Hi - I think your argument here about still needing EV because accounts may be done on a prudent basis is based on an assumption that Solvency II balance sheets are only available privately to the regulator. But they are public disclosures - so analysts etc can see the value of "own funds" for a company. Hence the arguments made in the Core Reading about EV having reduced relevance.

So, in the example above, a company ends up having to do two sets of liability calculations and potentially two sets of EV calculations (if the second EV mentioned above is required). It seems like far less work to just align the accounts to reflect SII liabilities (if these can be seen to be more relevant and no less reliable). Would you agree?

Yes! Hence some companies have moved from Solvency I to Solvency II liabilities in their accounts. However, those who are subject to IFRS may choose to focus on the transition to IFRS 17 (which is a lot of extra work for insurers) rather than also having this intermediate change.

 

[Lindsay Smitherman]

Please can you confirm my understanding of the following

Yes - what you have written here seems broadly OK, although bear in mind that there is still some flexibility allowed in terms of what particular companies actually do in practice.

How does the company decide when to include the latter?

The RC must include any capital whose distribution to shareholders is restricted in some way. It is up to the company to assess whether or not it has any such restrictions in place above and beyond its regulatory capital requirements.

(2) is just the value of these capital amounts at the valuation date - ie no need to project them and discount the releases because the discount rate and investment return assumptions are the same in MCEV

Yes, but note that (2) is always just the value of those capital amounts at the valuation date, irrespective of whether EEV or MCEV. If using a traditional EEV, with a risk margin loaded into the discount rate, then the impact of the difference between the (higher) discount rate and earned rate falls into (3) ie the cost of lock-in.

(3) requires all the elements of (2) to be projected so that a chosen cost of capital rate (not necessarily 6%) can be applied to those projected capital amounts (and then discounted). This would be the frictional cost of capital

The frictional cost of holding locked-in required capital doesn't have to be determined using a "cost of capital" approach as such. Some companies could explicitly calculate the frictional costs, eg the additional tax and investment expenses on the amounts involved.

 

[Mbotha]

Lindsay, you are amazing! Thank you for the help!!

 

[WanderingEejit]

Lindsay, please excuse me butting in here. I still can't get my head around one of the (apparently) simpler issues above. Why is PVIF zero?
Assuming no matching/volatility adjustment or contract boundary issues, won't the cashflow projection used to derive the BEL have offsetting positive cashflows arising from charges, leading to a positive VIF? For example on unit linked pension business, I can understand VIF being zero if PV charges equalled PV expenses plus benefits (etc), but the profit criterion at time of pricing would have assumed PV of charges would be greater (above some hurdle).
Furthermore, on certain existing business the historic experience will be such that there is even bigger profit to be gained over future years than expected at time of pricing - eg if fund sizes are now much larger than were expected at time of pricing, PV of AMC income on unit linked pensions will be even higher than (PV outgo + NPV equivalent of profit requirement).

Hope this makes sense. Just can't reconcile what I'm seeing in work with this part of the notes!

 

[Lindsay Smitherman]

Hi - it's because we now have the balance sheet presented on the Solvency II basis and the liabilities are therefore best estimate rather than prudent. When liabilities used to be prudent, VIF represented the expected future emergence of the prudential margins. These margins fall into profit, provided that actual experience = best estimate. No prudential margins in liabilities -> no VIF.

What is happening is that all of the profit loadings within a product are capitalised (i.e. taken fully into account in the EV calculation) when the business is initially written. In your alternative example, where higher profits arise due to better than expected historic experience, these additional profits are fully recognised when that better experience happens.

In the UL pensions example that you have given, the fact that charges are higher than expenses etc means that the non-unit part of the BEL is negative. Hence free surplus is higher than otherwise, and that is how those inherent profit loadings are taken into account in the EV.

So the EV is higher than it would otherwise be, but effectively the present value of future profit margins is found within the free surplus (or net assets) part of the EV, not the VIF. (Unless, as described in the course notes and as you indicate, there are profits emerging beyond contract boundaries or other expected profits which the BEL does not allow us to recognise.)

Does that help? Hope so!

 

[Lindsay Smitherman]

Lindsay, you are amazing! Thank you for the help!!

You're very welcome!

 

[WanderingEejit]

OK I think there's a disconnect in my thinking with the VIF vs 'change in value of existing business' (ie AoS items). So am I right in thinking the PV of future charges will be recognised as profit when the new business is written and added to free assets in that year, with the result that in all future years there is no profit or loss on that business (under SII margin-free basis) other than AoS experience items (which again increase or decrease frequency surplus)? So in this sense there is no VIF on any business, though at the point of writing new business there is really some VIF (ie positive present value of contract) - it just happens to be recognised through AoS for that year?

 

[Lindsay Smitherman]

Yes!

And if there is no VIF (no contract boundaries, no illiquidity premium etc) then the analysis of change in embedded value just becomes an analysis of surplus.

 

[WanderingEejit]

Yes!

And if there is no VIF (no contract boundaries, no illiquidity premium etc) then the analysis of change in embedded value just becomes an analysis of surplus.

Brilliant, thank you Lindsay. This makes much more sense to me now. Really appreciate the fast replies