Cost of smoothing & guarantees

[Mbotha]

I have some questions on the definitions of these:

• Cost of guarantees = max(0 , guaranteed benefit - AS)
  • Why is this sometimes calculated using smoothed AS?
• Cost of smoothing = payout - max(AS , guaranteed benefit)
  • There would be a cost if the payout (on death or maturity) is based on smoothed AS rather than unsmoothed (or base) AS, where smoothed AS results from either smoothing the AS itself or from smoothing the investment return used to calculated the AS. Is that right?

Any help would be appreciated. Thanks!

 

[Lindsay Smitherman]

Hi. These concepts don't have formal definitions, so try not to get too bogged down in them. Have you perhaps found these from the April 2014 SA2 exam paper? That question was about the Solvency I Peak 2 realistic balance sheet calculations, which are no longer in-force for supervisory reporting due to the introduction of Solvency II. Under those calculations, the cost of guarantees and cost of smoothing had to be shown as separate components on the supervisory balance sheet - but this is no longer the case and you don't need to know these details.

Also, the question included these particular definitions as what the particular company in the given scenario was using to perform the calculations. A company may split out the difference between unsmoothed asset share and payout in different ways. As for an analysis of surplus, it depends on the order in which the analysis is done. Here, for example, whether:

Unsmoothed asset share -> impact of smoothing -> does the guarantee bite -> payout
or
Unsmoothed asset share -> does the guarantee bite -> impact of smoothing -> payout

Your definitions take the second route, assuming that cost of guarantee is based on unsmoothed asset share. Alternatively, the first route would calculate the cost of smoothing component as smoothed asset share minus unsmoothed asset share, and then the cost of guarantee component as payout - max(guaranteed benefit, smoothed AS}.

Payouts are normally based on smoothed asset share. And yes, smoothing can be done in different ways by different companies, eg through smoothing the investment returns applied to the asset share calculation, or by smoothing or limiting the overall % change in asset share from period to period, or by basing terminal bonus rates on unsmoothed asset share and then smoothing the bonus rates over time, etc.

 

[Mbotha]

Have you perhaps found these from the April 2014 SA2 exam paper?

Yes, I did. Thank you for clarifying and for the details.

 

[User 1234]

Hi. These concepts don't have formal definitions, so try not to get too bogged down in them. Have you perhaps found these from the April 2014 SA2 exam paper? That question was about the Solvency I Peak 2 realistic balance sheet calculations, which are no longer in-force for supervisory reporting due to the introduction of Solvency II. Under those calculations, the cost of guarantees and cost of smoothing had to be shown as separate components on the supervisory balance sheet - but this is no longer the case and you don't need to know these details.

Also, the question included these particular definitions as what the particular company in the given scenario was using to perform the calculations. A company may split out the difference between unsmoothed asset share and payout in different ways. As for an analysis of surplus, it depends on the order in which the analysis is done. Here, for example, whether:

Unsmoothed asset share -> impact of smoothing -> does the guarantee bite -> payout
or
Unsmoothed asset share -> does the guarantee bite -> impact of smoothing -> payout

Your definitions take the second route, assuming that cost of guarantee is based on unsmoothed asset share. Alternatively, the first route would calculate the cost of smoothing component as smoothed asset share minus unsmoothed asset share, and then the cost of guarantee component as payout - max(guaranteed benefit, smoothed AS}.

Payouts are normally based on smoothed asset share. And yes, smoothing can be done in different ways by different companies, eg through smoothing the investment returns applied to the asset share calculation, or by smoothing or limiting the overall % change in asset share from period to period, or by basing terminal bonus rates on unsmoothed asset share and then smoothing the bonus rates over time, etc.

Hi Lindsay,

Can I please get more clarity regarding to the 2014 Apr SA2 exam?

The solution shows the detailed calculations of each component AS/Smoothed AS/Guarantees, then perform comparison for the case of Maturity claim. On the other hand, when it comes to the case of Withdrawal claim, it gets briefly mentioned without performing calculations (then comparisons) as below:
1. There is no costs of guarantee for withdrawals as they received the un-smoothed asset share.
2. There is no costs of smoothing for withdrawals as they received the un-smoothed asset share.

For 1, I can understand it because that they received un-smoothed asset share (i.e. no smoothing applied at all), in such case there is no point to perform the calculation of costs of smoothing.

However, I'm struggling to understand why don't we need to perform the comparison between un-smoothed AS with guarantee (although the numbers in the question indeed returns zero)?

Thanks in advance!

 

[Lindsay Smitherman]

As stated in the question, the surrender value equals the unsmoothed asset share.

The guaranteed benefit that underpins with-profits business normally only applies to 'contractual benefits', ie on maturity and death. Surrendering a policy before the guarantee 'breaks' the contract and the guarantee would not apply.

The fact that the MVR is currently in place reinforces this idea: the company is currently paying out less than the unit value on surrenders. [It would not be able to do this at maturity, as the guaranteed minimum payment of the unit value would apply at that date.]

If the company is always paying out the asset share on surrender, there is no cost of guarantee on surrenders.