Key findings from the consultation on the Draft Solvency Assessment and Management Second South African Quantitative Impact Study (SA QIS3) Technical Specifications

yellowgreatΤεχνίτη Νοημοσύνη και Ρομποτική

16 Οκτ 2013 (πριν από 3 χρόνια και 10 μήνες)

189 εμφανίσεις

Key findings from the consultation on the Draft Solvency Assessment and
Management Second South African Quantitative Impact Study (SA QIS3)
Technical Specifications


Page
1

of
176

Introduction

The draft SA QIS3 technical specifications were published for comment on the 2
nd

of August, and the consultation
was open for comment until the 30
th

of August. There was a strong response to the call for consultation, with a total
of 463 comments received from 26 respondents.

The FSB went through the comments and allocated these to either themselves or to the relevant Task Groups to
respond to. The dr
aft responses to the comments were then tabled at the Pillar 1 Sub Committee held on the 17
th

of
September. The discussions held at the Pillar 1 Sub Committee have been used to finalise the responses to the
comments as well as the changes required to the t
echnical specifications.

Instead of tabling all 46
2

comments and responses at the SAM Steering Committee, a summary of the

key changes as
set out in the following pages of this

document
were tabled
as a basis for discussion.

This document contains the
summary presented at the SAM Steering Committee (pages 1 to 5) as well as the full list
of comments received, with written responses and resolutions (pages 6 to 176).


Technical Provisions

Risk Free Rate

The draft SA QIS3 specifications required the use of

the government bond rate as the basis for the risk free rate.
Based on the feedback received, this will now be changed such that insurers may use the swap rate to discount the
cash flows of liabilities that are backed by assets invested in swap based asse
ts. This will mean that there is less
volatility as assets and liabilities will be valued consistently.

There will be additional qualitative questions to understand how the choice of risk free rate can be linked to the
investment policy of the insurer.

Ill
iquidity Premium

There were some comments received requesting that an illiquidity premium be added to the risk free discount rate.
However, given the results of previous SA QIS exercises, the default position will remain that there is no illiquidity
premiu
m to be added.

Insurers are invited to provide additional information on what they think the illiquidity premium should be, and what
the impact of an illiquidity premium limited to 50 basis points would be.

Segmentation

There were some comments received qu
estioning the segmentation put forward in the draft SA QIS3 specifications,
specifically where there were some differences between the draft specification and Discussion Document 29 v5,
dealing with the segmentation for authorisation. A few small changes w
ill be made to sub classes to allow better
alignment.

Key findings from the consultation on the Draft Solvency Assessment and
Management Second South African Quantitative Impact Study (SA QIS3)
Technical Specifications


Page
2

of
176

There were further comments requesting clarification on how products should be segmented, and how products
should be unbundled between different classes.

Taxation basis

Clarification was requested on
the tax basis to use. It was clarified that the current (SVM) basis should be used for
determining tax.

Risk Margin

There was a comment on whether it is prudent to assume loss absorbency of deferred tax assets and deferred tax
liabilities for the SCR which

is used to calculate the risk margin. It was agreed that this loss absorbency should be
removed for the purpose of the risk margin.


SCR


Market Risk

Credit Risk

It was noted that there was no table mapping from local credit rating to international
credit rating. This is still under
development by the Working Group and will be released as an annex to the tech spec as soon as it is complete.

In general, there are many comments relating to credit risk which are answered in Discussion Document 99 being
developed by the Credit Risk Working Group.
Responses refer to this

draft Discussion Document
.

Concentration risk

There were a number of comments stating that the formula for concentration risk was onerous, given that there are
only 4 large banks at which
insurers would spread deposits. This concern is noted but there are no changes to the
specification as the concentration risk is still present and should not be ignored.


SCR


Life Underwriting Risk

Lapse Aggregation

A simplification from the approach tested in SA QIS2 for aggregating a mass lapse event with a change in the level of
lapses was proposed in the draft SA QIS3 specifications. However, there were concerns with the onerousness of the
simplification. In addit
ion, there were comments that lapse risk should not be simplified, as it produces one of the
highest capital requirements of all the sub modules.

The updated SA QIS3 specification is now in line with the method tested in SA QIS2.

Expenses on mass lapse sc
enario

Key findings from the consultation on the Draft Solvency Assessment and
Management Second South African Quantitative Impact Study (SA QIS3)
Technical Specifications


Page
3

of
176

There were concerns that by including expenses within the mass lapse scenario, this would lead to a double counting
of expense risk as it is already allowed for in a separate sub module. However, it has been decided to stick to the
methodology set f
orward in the draft SA QIS3 specification. Further work will be carried out to understand whether
or not there is double counting of expense risk.


SCR


Non
-
life Underwriting Risk

Parameters

There were a number of parameters that were questioned by stakeh
olders. The Non
-
Life Underwriting Task Group
considered the comments received, and the parameters have remained largely unchanged.

There were also comments received questioning whether the Non
-
life underwriting risk component should allow for
the expected
profit that an insurer would make over the next year. It has, however, been agreed that the expected
profits should not be allowed for in the non
-
life underwriting risk component of the SCR.

Non
-
Life Lapse

Based on comments stressing how immaterial the non
-
life lapse stress is, especially due to the short contract
boundary, it has been agreed that the non
-
life lapse stress may be excluded if it is immaterial.

SCR


Other

Participations

There were 3 different methods for allowing the risk charge for strategi
c participations in the SCR that were put
forward in the consultation. They were:

Method 1


Account for the participations as equity, but allow no diversification with any other risk charge. This is
what was tested in SA QIS2, and is the default for SA QI
S3.

Method 2


Base the charge on a look
-
through approach, whereby the charge is calculated based on the own funds
and the SCR of the participation.

Method 3


As per Method 1, but allowing for diversification

There was no clear favourite based on the comm
ents received. The approach for SA QIS3 thus remains that Method
1 will be the default approach, but will also ask for results on the other approaches. There will be the following
changes:



For Method 1, the lack of diversification benefit will only be for
insurance or insurance related participations



For Method 2, approximate values can be used where the insurer does not know the exact values required
for participations.



Method 3 will be shown for illustration purposes in the SA QIS3 Core Workbook.

Key findings from the consultation on the Draft Solvency Assessment and
Management Second South African Quantitative Impact Study (SA QIS3)
Technical Specifications


Page
4

of
176

Ring
Fenced Funds

Concerns were raised that the approach for cell captives was onerous, and not well suited to the nature of cell
captives. Some further information will be requested at the cell level so that further options on the allowance of
diversification
across cells can be tested.

Concerns were also raised on whether discretionary participation funds should be treated as ring fenced or not.
Further work will also be conducted on this.

Gross and net of management actions

The requirement in the draft SA QIS
3 technical specifications to produce the results of stresses both gross and net of
management actions has been removed. This has now been replaced with a requirement to disclose the position of
the with profit funds based on the disclosures put forward in

the draft Life Technical Provision QRTs that were
consulted on earlier this year.

Double counting of management actions

The weighting to be used for the double counting of management actions will be changed from the SCR impact to
the asset impact. This wi
ll better capture the potential double counting of management actions where there the
management actions result in a very small or zero SCR after a stressed market risk event.

Loss absorbency of Deferred Tax Asset

Whereas the loss absorbency of deferred ta
x liabilities is relatively well understood, there were questions on
whether, and if so how, the loss absorbency of setting up deferred tax assets should be allowed. It was agreed that
the loss absorbency of deferred associated with setting up deferred tax

assets may be allowed, with some further
guidelines given. Further qualitative questions will be added to understand how insurers have gained comfort that it
is appropriate to set up deferred tax assets.

Liquidity assessment

There were concerns that the l
iquidity assessment put forward in the SA QIS3 technical specifications would be too
onerous, thereby making the assessment obsolete as it would not be a good indicator of whether an insurer may
have liquidity issues after a stressed event. It was agreed t
hat additional information would be requested, as derived
by the Task Group to give further input as to what may provide a good liquidity risk indicator.

Own Funds

Treatment of own shares

Further clarity was required on whether own shares should be
excluded from own funds where the policyholders are
exposed to the risk associated with the shares. SA QIS3 will be updated to allow insurers the option to apply the
delta NAV approach to own shares where they would want to include own shares held due to t
he risks associated
with the shares being borne by policyholders.

Key findings from the consultation on the Draft Solvency Assessment and
Management Second South African Quantitative Impact Study (SA QIS3)
Technical Specifications


Page
5

of
176

Limit that insurers may hold in a bank in the same group

There were a number of comments received on the new proposal limiting the amount of own funds that the insurer
may take into account
from deposits in a bank that belongs to the same financial conglomerate. The concerns were
noted, however no change has been made to the requirement. One specific concern was in relation to the significant
impact that this may have on some insurers. Care w
ill be taken when interpreting the SA QIS3 results for these
insurers.

Removal of PU GAP calculation

Based on comments received, the requirement to calculate the paid
-
up gap calculation has been removed.
Furthermore, the surrender value gap has been modifi
ed to allow for the clawback of any commission in event of a
surrender.

Groups

Clarification on alternatives to be calculated

The draft SA QIS3 specifications included the following calculations:



Deduction & Aggregation using current capital requirements



SA QIS3 Deduction & Aggregation



SA QIS3 Accounting & Consolidation

The SA QIS3 specifications have been updated to state that the first two calculations are required, and that the third
calculation is optional.

Removal of combined approach

The draft SA QIS
3 specifications included specifications for a combined approach, using a combination of the
deduction & aggregation approach and the accounting consolidation approach. This has now been removed as this
calculation is not required for SA QIS3 approach.

Exc
lusion of non
-
equivalent subsidiaries

The draft SA QIS3 specification had not fully been aligned with the latest position papers from the Insurance Groups
Task Group. Specifically, the concept of excluding subsidiaries from countries that are not likely to

be deemed
equivalent as set out in Position Paper 85 was not allowed for in the draft SA QIS3 specifications. The specifications
will be updated to allow for this.

More generally, the specifications will be checked for consistency with the Insurance Group
s position papers.


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
6

of
176

No

Section or paragraph
reference

Comment

Response

Proposed changes

1

Covering Letter.

Risk Free Rate

It is preferable to use the government bond rate as the
basis of the risk free rate.

See response to comment 393

None

2

General


Throughout the
technical specifications QIS2 is sometimes
referred instead of QIS3


Noted

To be corrected.

3

General

Bond/Swap: No preference

Noted.

None

4

General

Contract Boundaries: Not material for non
-
life insurers

Noted.

None

5

General

Lapse Risk: Relevant to
life insurers only

Noted

None

6

General

Noted several references to QIS2 when it should have read
QIS3


suggest find and replace (with care) (eg Part 1 pg 3,
eg SCR11.1.1)

Noted

To be corrected.

7

General

Participations

1.

Fair value/No diversification:
This approach is
overly conservative

2.

Look through: This approach poses a timing
problem in that the SCR of the participation
must be known before finalising the SCR
calculation for the owner. This approach is
rather impractical

3.

SAQIS1: This is our preferre
d approach. A
diversification benefit does exist and so
1) This argument is addressed in the
discussion docu
ment, DD53 and will
be subject to further debate.


2) The principle that applies is not
much different from the current
regulatory regime requirements.
However we do concede that the
greater complexity of SAM (versus
current regime) as well as the
shorter

reporting periods that will
apply under SAM will make this a
No
change required


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
7

of
176

should be allowed. A correlation matrix
between equities and participations can
determine the extent to which a
diversification benefit is allowed.

greater challenge going forward.


3) We take note of preference. The
final method for participations has
not been finalised and this is why
more than one approach is being
explored in QIS3.

8

General

Sections highlighted in yellow in this document are in
response to the questions highlighted in yellow in the
covering letter.

Noted

None

9

General

All Sections

There are a number of spelling errors so a general spell
check is
needed. There are also a number of references to
QIS2 which should read QIS3 e.g. TO29.9, OF8.3, SCR9.3.1,
SCR10.2.12, SCR11.1.1, SCR11.4.1, SCR 12.1.1, SCR 14.2,
SCR14.2.1, SCR14.2.2, IM1, OF8.3.

Noted

To be corrected.

10

General

Overall comment

Where
the final SAM basis differs from discussion
document recommendations these should all be updated to
document the final basis and rationale for the change, e.g.
risk free rate moving from swaps to the government bond
curve and removal of the illiquidity mat
ching adjustment.


Noted.


No change required.

11

General.

Prescribed yield curve

Irregular forward rates


We propose that any prescribed yield curve to be used at
the various dates does not have irregular forward rates (i.e.
smoothed to an extent)

Valid comment, see response to
comment 182.

None


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
8

of
176


There

could be significant discontinuities in both the
nominal and real forward rates. This appears to be the
result of deriving the rates by fitting exactly to market data
points. These discontinuities do not reflect true differences
in the implied future inte
rest rates rather they reflect small
inconsistencies between market data points.


We would prefer the derivation of the interest rate curves
to strike a balance between achieving a close fit to market
data and achieving smooth interest rate curves.


The as
set model used to value embedded options and
guarantees is calibrated to the nominal and real interest
curves. It is preferable to calibrate the model to smooth
interest rate curves. The use of non
-
smooth interest rate
curves can lead to the model producin
g asset returns that
are not market consistent and risk
-
neutral.


12

Introduction

Please provide the final Excel return templates when the
final specifications are issued as it will help to interpret the
specifications.

The Excel return templates will go
through a process of development
checking and testing for quality
purposes. Thus the expected release
date of the Excel return will be the
18
th

of October.

None


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
9

of
176

13

Introduction

The yield curve and other
parameters need to be available
very early in January 2014 in order to use 31 December
2013 as reporting date.

We currently receive yield curves and parameters on the
first working day of the month. A process should be put in
place to have the SAM curves a
nd parameters available by
the first working day of each month.

See response to comment 182.

None

14

Introduction

There are still references to SA QIS2 (which should be SA
QIS3) in the third and second last paragraphs of the
introduction.

Noted

To be
corrected.

15

Introduction page 3

Two references to “QIS2”. These should presumably be
“QIS3”?

Agree

Technical Specifications
amended to reflect
“QIS3”.

16

IRM

SCR.12.5.1

Part 2 of 6

Page 52

Approved / Non Approved Reinsurance is not taken into
consideration.

For SA QIS3 it is. The outcome of the
Reinsurance Review Project will
determine the treatment under SAM
once effective.

None.

17

LIQ

Granularity of liquidity assessment


We propose that a more granular calculation might also
assist in
making this calculation more reflective of reality


The Liquidity calculation is aimed at
obtaining an indication of the
potential liquidity strain of an insurer
after a stressed event. In
determining the calculation, a
balance is required between a
detail
ed calculation and a pragmatic
approach. The current calculation is
a first attempt. The information and
feedback from this calculation will be
used to determine whether a more
detailed or more pragmatic approach
No change required


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
10

of
176

would be appropriate.

1
8

LIQ.1

Liquidity risk

An alternative suggestion would be simply to compare the
worst (across all underwriting modules) after stress one
-
year liability cash flow with the liquid assets, after the worst
market SCR shock. This should only be done for
non
-
linked
business.

Noted. However, this may
understate the liquidity strain from
the liability side as it ignores all other
cash flow strains from other liability
stress events

No change required

19

LIQ.1

Liquidity risk

The section seems overly complex
and fairly conservative
for most insurance companies that clearly have no liquidity
constraints. Please send out the return template that does
this calculation with the specifications as this will help to
understand the requirements.

Noted. Please see
response to
comment 25

No change required

20

LIQ.1

Part 2

Liquidity Risk
Assessment

There is a requirement for a complex liquidity shortfall
calculation


i.e. compare liquid assets after a stress event =
SCR to the cash flows that would emerge over the y
ear
(excludes any investment return). We question the value of
performing this calculation


it is very onerous to perform,
and even then requires many assumptions to be made.

The calculation is surely only relevant to risk products and
all investment
products and annuities should be excluded.
For Fund business, there is also a typical provision that
assets can be transferred to the client rather than cash,
particularly on surrender. For risk products the capital over
1 year relates mainly to the mass l
apse and expense shocks
for life insurers. A simpler calculation just looking at these 2
The calculation has been designed
such that no additional runs would
be required to

obtain the inputs. As
there should already be runs
performed to produce the stressed
liabilities, the cashflows under these
stresses should be available.

Not all investment and annuity
products have no liquidity issues


especially where the assets and th
e
liabilities are not matched. However,
see response to comment 23 to
decrease the scope of the
calculation.

See also response to comment 17 on
the balance between a detailed
approach and a pragmatic approach
No change required


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
11

of
176

shocks would make more sense for life insurers. In detailing
the calculation the timing of cash flows must be specified as
this can materially impact the answer.

An as
sessment of liquidity risk is part of the ORSA. We
recommend that all insurers should use a few consistent
scenarios that are specified by the FSB to assess liquidity
risk.

to the calculation.

21

LIQ.1 p70 Part 2

Liquidity Risk Assessment

The changes that have been made to the Liquidity Shortfall
assessment methodology in the draft QIS3 specification
(when compared to the method proposed by the working
group in Discussion Document 95) invalidate th
e
methodology:

The draft QIS3 method now compares:



the available liquid assets following a fall in value
of all asset classes according to the SCR market
stresses, to



the cash outflows over the next year following a 1
-
in
-
200 year insurance risk event (with
out any fall in
asset values).

This creates a significant error which is easiest to
demonstrate with an example (not realistic but shows the
issue):



Assume the mass lapse stress was 70% and the
equity stress was 50% and there are no other risks.



Assume a c
ompany which only sold single
premium, listed equity, unit linked (not pure
Disagreed.

The proposed approach to use t
he
impact of an asset stress on both the
assets and the liabilities does not
work in all cases.

Consider an annuity portfolio where
the liabilities are backed by
government bonds. If there is a rise
in interest rates, the market value of
the assets will de
crease immediately.
Although this will be offset by the
change in present value of the
expected annuity payments, the
actual cash flows for the annuities
will only take place at a later point in
time. This may create a liquidity
strain if the assets need t
o be sold
now to meet other cash flow
payments.

The example highlighted is
addressed by the response to
comment 23.

Also note response to comment 25
which acknowledges the level of
The technical
specification will be
updated to also collect
information
as
requested under
Discussion Document 95


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
12

of
176

linked) business had a unit liability of 100
(technical provisions are irrelevant for this) and
assets of 120 invested in matching listed pure
equity at the valuation date. Clearly

this company
has no liquidity risk as all its assets are liquid and
the highest possible cash
-
outflow would only be a
portion of its assets under all possible scenarios.



Applying the method in the draft QIS3: The
Available Liquid Assets would be 60 (i.e.
120*(1
-
50%)) and the Cashflow Requirement would be
greater than 70 (i.e. 100*70% plus expected
cashflow)



The draft QIS3 methodology thus shows a
company with no Liquidity Risk as having a
Liquidity shortfall.

The reason the draft QIS3 methodology does not
work is
that it is comparing assets under market risk scenario to
cash
-
flows which have not been subjected to the same
market risk scenario. The comparison of unrelated
scenarios will give misleading results.

We see two potential ways of correcting the dra
ft QIS3
methodology:

1.

The assessment of the Cashflow Requirement
could be performed under a consistent market risk
scenario (i.e. all the calculations NetCFexpected
and NetCFstress

i should be performed assuming
all asset stress events as set out in the market risk
section of the SCR have occurred simultaneously);
or

2.

The assessment of the Available Liquid Assets
could be assessed NOT after the stress events as
set out in the market
risk section but rather
adjusted for the SCR (or delta BOF) of the relevant
(i.e. for liquid asset classes) market risk sub
-
prudence in the calculation.

However, it is noted that the
calculation put
forward in Discussion
Document 95 may provide additional
information. The technical
specification will be updated to also
collect information as requested
under Discussion Document 95


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
13

of
176

modules.

The first method would require significant additional
calculations as each of the insurance risk sub
-
modules
would need to
be recalculated assuming all market risk
stresses happen simultaneously.

The second method allows for the impact of a fall in assets
approximately without the need for additional calculations
and is consistent with the original proposal in the
Discussion D
ocument 95.

It is critical that this error is corrected to ensure valuable
information on this risk is assessed as part of the QIS. Our
preference is to follow the method consistent with the
Discussion Document 95.

22

LIQ.1.4

Does this assumption allow for the impact of the fire
-
sale of
assets of all insurers in the market given that the market
risk stresses are industry wide, which will decrease the
value of assets even
further? (pro
-
cyclicity)

This calculation does not take the
pro
-
cyclicality effect into
consideration. However, the pro
-
cyclicality would only be relevant if
all insurers were required to sell
assets, which is not the case

No change required

23

LIQ.1.5

Liquidity calculation


classes of business included


It should be considered to also exclude cash flows and
assets held in policyholders’ unit accounts such as unitised
with
-
profit and universal life, only including non
-
unit cash
flows and assets or exclu
ding all business where there are
not significant life underwriting SCR (but including an
Agreed. However,
will add the
stipulation that this is only the case
where the assets held match the unit
liabilities. If this is not the case, there
would still be a liquidity strain.

Changes to be made to
reflect comment


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
14

of
176

assessment of the non
-
unit flows under a mass lapse
event). This is suggested in order to make the liquidity
assessment more reflective of reality.


24

LIQ.2.2

Liquidity risk

In the calculation of
Available liquid assets, it should be
made clear how to treat shocks that have different
directions (e.g. interest rate shocks) and (e.g.) whether it is
necessary to apply currency shocks over and above equity
shocks for offshore equity.

Agreed.

For shocks

that have different
directions, the most onerous
direction on the available liquid
assets should be used (regardless of
what shock is more onerous in the
SCR calculation)

Currency shocks should be applied in
addition to the stresses.

LIQ 2.2 to be amended

to clarify treatment

25

LIQ.2.2

Is the level of prudence excluding the diversification benefit
not so large as to render the assessment invalid?

The liquidity assessment has no
impact on the SCR and hence on the
solvency of the insurer.


The approach taken is to make the
calculation overly prudent, as there
would be no negative impact on the
solvency. Insurers typically have a
strong liquidity position, so it is
expected that even with the prudent
assessment, most insurers should
show enou
gh liquidity. Due to the
prudence, a liquidity shortfall will not
necessarily mean that there are
liquidity issues, but would prompt
further, potentially more detailed
assessment of the liquidity position
No change required


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
15

of
176

of the insurer.

26

LIQ.2.2

No
mention is made of whether the aggregation matrix
should be applied.

The formulae in LIQ.2.4 includes the
Correlation matrix.

No change required

27

LIQ.2.2

Liquidity risk

The specification should clarify how Cashflow@risk

should
be calculated for modules that have sub
-
modules, e.g.
Lapse risk.

Agreed.

For the lapse stress, the mass lapse
scenario should be considered

Specification to be
updated to include
clarification

28

LIQ.2.2 & LIQ.2.4

These sections refer to the calculation of cashflows in
respect of the life and non
-
life underwriting risk modules.
However it is not clear why the non
-
life underwriting risk
module is referred to because in the preceding section LIQ
1.5 the specifications

specifically exclude non
-
life insurance
from the liquidity calculation.

Agreed.

References to non
-
life in
the following paragraphs
to be removed: LIQ.2.2,
LIQ.2.4,

29

LIQ.2.4

Why are the non
-
life stresses included if the stress only
applies to life compa
nies as per LIQ 1.5?

Refer to comment 28.

Refer to comment 28.

30

MCR.2.6

“The values specified in i and ii above will de adjusted
annually to ensure that they do not reduce in real terms.”


this reference is invalid

Agreed, this is relevant for the
discussion document and not for the
QIS3 specification where the value is
specified for the exercise. The
paragraph will be deleted.

Delete paragraph.

31

MCR.2.6

Assume “values specified in i and ii” refers to (a) and (b) in
MCR.2.5?

Agreed.

This text
will be removed
for the final QIS3
specification.

32

MCR.2.6

Should read “The values specified in (a) and (b) above will
be adjusted...” there’s no i and ii above.

Agreed.

MCR2.6 will be removed
in the final spec.


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
16

of
176

33

MCR.4.5

Part 2 of 6

Page 64

It
appears that the table for α and β has not been updated
with SA data and therefore remains non
-
reflective of SA
risks.

Further: Percentages for additional LOBs appear to be
aligned to the original Miscellaneous class


We received no objections to the
param
eterisation proposed in SA
QIS2. Also, the results from QIS2
were assessed to be reasonable in
the SA context

Given the comments above, the
parameters were adjusted to align
with the updated segmentation for
QIS3.

None

34

MCR.4.5

It seems that the
previous factors that were used for
Miscellaneous were used for all the lines that used to be
part of that class. This result in classes like “legal” and
“travel” having higher factors than “engineering”. Does this
make sense? (Updated discussion document
available yet?)

The results from QIS2 were assessed
to be reasonable in the SA context.
Therefore, the parameters were
adjusted to align with the updated
segmentation for QIS3.

None

35

MCR.6

Are composites going to be allowed? Can this section be
removed?


No composite method should be
used. Current composite insurers
should complete two returns


one
for the LT business and one for the
ST business


Remove any reference to
the composite method.

36

MCR.6

MCR

Part 2 of 6

Part 68

We request the FSB to
provide guidance on how composite
licences, which are the majority of Reinsurers, should
approach MCR 6 and the SCR calculation in general. Should
the composite method be followed or should a separate
approach for Non
-
life and Life be followed?


No composi
te method should be
used. Current composite insurers
should complete two returns


one
for the LT business and one for the
ST business


Remove any reference to
the composite method.

37

NLUR

The std

deviation for Miscellaneous: Warranty appears too
Noted

None since it has been
decided to merge all
Miscellaneous sub
-

Composite

comments
for SA QIS3 Draft Technical Specifications


Page
17

of
176

Part 5 of 6

SCR.8.2.18

Page 12

high.

classes in SA

QIS3 in line
with DD29.

38

OF.12.2

Tiering of hybrid capital and subordinated debt


We believe that the criteria of
tiering of hybrid capital and
subordinated debt as specified in this paragraph should be
relaxed so that it is no longer allocated to tier 3 in its
entirety, but rather follows the rules in the remainder of
the Own funds section.


Agree. The QIS3 tech
-
spec
s will be
updated.

Tech specs to be
updated

39

OF.6

Own Funds

Part 1 of 6:

Page 101

It is unclear from the section that Intangible Assets should
be treated as Tier 1. Kindly add more guidance.

The intangibles (after the deduction
set out in OF.6) should
be treated as
Tier 3 own funds

Clarification to be
included

40

OF.7

Part 1

Own Funds

Further deductions from
Basic Own Funds BOF

A clear definition of non
-
linked assets is required to ensure
exclusion of any funds managed on behalf of shareholders
in
calculating the spreading requirements.

Agreed

Further guidance to be
provided in section OF.7


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
18

of
176

41

OF.7

Holding companies included


Holding company could possibly include holding company
of a holding company of an insurer. Not sure if clarification
is
provided elsewhere. Reference could be made to
“controlling” and “non
-
controlling” holding companies.


The condition applies to any holding
company, whether controlling or not

No change required

42

OF.7

(Basic Own Funds)

The proposed deductions from Basic

Own Funds are severe
for life insurance operations that form part of the main four
banks, especially in respect of the 10% limit on deposits.

The straight deduction as opposed to a demotion of the
assets to a lower tier seems overly severe. Solvency II ev
en
allows for off balance sheet items to be considered by the
regulator for inclusion as well as the addition of
subordinated debt to basic own funds. We propose aligning
SAM to current admissible asset limits (e.g. 20% limit on
deposits) and to have holdi
ngs in excess of the limits
demoted to lower tiers of own funds instead of a straight
deduction.

See response to comment 45


43

OF.7.3

Spreading requirements
-

Holding company shares


Listed shares in holding company exclusion should only
apply to internal funds where the liability would not change
Agreed

Clarification to be
provided


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
19

of
176

consistently with the change in asset value. Also, where a
large part of the market risk is passed to policyholders e.g.
discretionary parti
cipation business, clarity is also required
around the treatment. For instance, reference could be
made to policyholders’ unit accounts rather than to
“linked” or “unit
-
linked” business.


44

OF.7.3

Page 102

The 5% limit

for financial conglomerate is not supported as
there are very limited options in South Africa to invest
funds in a secure institution. We strongly recommend that
this is reconsidered. By investing funds with the holding
company is creates an implied paren
tal guarantee. The risk
for loss is extremely low. We are of the opinion that the
risks have already been adequately address in the SCR. This
is extremely punitive.

OF.7.3 only relates to investments
that the insurer holds in a holding
company of that insu
rer. There is no
mention of financial conglomerates
in paragraph OF.7.3

No change required

45

OF.7.4

Part 1 of 6, page 102.

“OF.7.4 Any cash balances held by an insurer at a bank
which is part of the same financial conglomerate in excess
of 10% of the
non
-
linked assets of the insurer should be
deducted from Basic Own Funds.”

The inclusion of this condition in the technical specifications
will adversely and unfairly (it is submitted), penalize
insurers affected by it


especially non
-
life insurers where
liabilities are largely backed by cash reserves. It appears
that the main reason for this insertion is to incentivize the
spreading and diversification of assets. However, this
‘charge’ is essentially already captured largely in the
concentration and credi
t risk components (SCR 6.7 and SCR
6.6 respectively) of the formula. Concentration risk, allowed
for under the market risk section of the standard formula,
shocks the excess of assets above a certain percentage of
Noted


The key reason for including
this condition is to prevent
concentration risk within a financial
conglomerate. This is separate to the
general concentration risk captured
within SCR 6.7. By including the
condition, the insure
r will be
disincentivised to keep large cash
balances within a bank that is within
the same financial conglomerate.
This would particularly be for when
the bank is in stress, and there may
be pressure from the bank to require
an insurer to hold more of its

cash
balances with that particular bank
No change required


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
20

of
176

total assets, according to the credit qual
ity of the asset. If
assets are relatively over invested in a certain instrument
this will increase the capital requirement accordingly
because of this excess concentration. Requirement OF.7.4
would seem to double count this source of risk
unnecessarily.

T
he economic impact of this on the industry may force the
funding of insurance companies to be taken away from the
banks due to this requirement. Our recommendation would
be to give due consideration to removing this condition in
its entirety, or, at the ve
ry least, revising it.


for liquidity purposes.

This is a particular issue for the South
African insurance market where
there is a large interconnectedness
between banks and insurers.

However, the comment is noted and
the FSB will consider

any changes
that may be required after analysing
the SA QIS3 results.

Care will be taken in interpreting the
results from this requirement to
avoid reaching incorrect conclusions

46

OF.7.4

Page 102

Own Funds

Part 1 of 6:

Clarity is
required what is meant with “cash balances”. Does
this only include cash balances or also bonds etc.?

This provision has been included due
to concerns where a bank requires
the insurer to hold its cash and other
short term deposits with that bank.

Clarification to be
provided that this should
only be applied to cash
and other short term
deposits

47

OF.7.4

Deductions to basic own funds

Whilst we recognise the need to avoid systematic risk
through O.F.7.4, this is challenging for some of our smaller
insurers e.g. in one of our small insurance companies we
have assets that are invested across the 4 banks. Notionally
we would then invest 25% in each of the four big banks. If
our insurer is regarded as part of the same financial
conglomerate as one of th
e banks, we would need to
reduce the cash in this bank to 10% and increase the
See response to comment 45

No change required


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
21

of
176

holding in the other banks to 30% each resulting in a less
diversified portfolio.

It it therefore worth considering whether the 10% should
be increased to allow for the small nu
mber of banks in the
SA industry.

48

OF.7.4

Definition of conglomerate


Should ideally define financial conglomerate. Suggestion is
possibly to include bank if it is considered in Group
submission


Agreed

Financial conglomerate
to be defined in section
OF.7.4

49

OF.7.4

(Strategic Participations)

It is unclear exactly what the definition of “cash balances”
is. If the term is interpreted broadly, it may be extremely
onerous to insurers who are members of ban
king
conglomerates in light of the limited number of investment
instruments available in the South African market.

Refer to comments 46 and 45.


50

OF.7.4

Refers to “cash balances” held. Should this be interpreted
as cash only, or include items such as
fixed deposits, term
deposits, preference share and/or debt instrument
investments, etc.?

Refer to comment 46


51

OF.7.4

Page 102

The current spreading limit is 20%. We would appreciate
more clarity in why lowering the current requirement.

See response to

comment 45

No change required


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
22

of
176

52

OF.7.4

Restrictions of own
funds to 10% cash
deposits in intergroup
entities.

This will reduce the own funds of insurance entities that are
owned by banks or are a part of a group with a bank as part
of the group.

The
risk inherent in these investments is not significantly
different because they are part of a group that contains a
bank.

See response to comment 45

None

53

OF.8

Are SV and PUP gaps still important given the proposed tax
basis? (as per our comment in
TP33.8.1 above)

Agree that PVG’s are not necessary
for QIS3. There is however still value
in disclosing SVG’s.

Remove the PVG
calculation but retain the
SVG. Allowance to be
made for items such as
“claw backs” in the
calculation of the SVG.

54

OF.8

SVG
and PVG

Are these meaningful? Will have to approximate these.

Refer to comment 53.


55

OF.8

Paid
-
up value gap


It is currently very difficult to determine a paid
-
up value
gap. Are there any approximations that could be used in
order to determine this gap relative to the surrender value
gap? Also, it would be helpful to understand whether
tiering requirements post S
AM implementation for this
item (as well as in respect of the surrender value gap) might
change.


Refer to comment 53.



Composite

comments
for SA QIS3 Draft Technical Specifications


Page
23

of
176

56

OF.8.10

p103 Part 1

Assessment of the surrender value gap

The surrender value gap as it is currently defined
significantly overstates the impact of a full surrender as it
ignores items such as commission claw
-
back which
significantly reduces the risk of surrender. As this value
does not appear to be used for anyt
hing this is not a serious
issue but should be highlighted to users of the information.
Alternately the formulae should be adjusted to allow for
these types of cashflows. In effect this would be similar to
the PVG assessment


i.e. compare the change in BE
L due
to a full surrender event.

Refer to comment 53.


57

SCR.

General SCR

The SCR formula, especially the market risk module, non
-
life underwriting risk module is becoming more and more
complex with each QIS and should be simplified.

Insurers should be
capable of determining their capital
requirements on a regular basis and need to be able to
report to the FSB on a quarterly basis. Given the
complexity, the time taken to gather all of the information
and perform the calculations will make this impossible

to
do so without significant simplification.

The skills required to be able to understand the outcome of
each risk module and sub
-
risk module are ever
-
increasing.
This is especially true of the market risk module and non
-
life underwriting risk module.

It
is impossible to find a standard formula that fits all

The current approach attempts to
strike a balance between simplicity
and adequately reflecting the
individual risk profiles of insurers.
The
economic impact study may
highlight any deficiencies in the
model. Insurers who are struggling
with their own implementation of
SAM should engage with the FSB.


None required.


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
24

of
176

insurers. A slightly conservative, simplified formula should
be chosen that will broadly ensure that the industry as a
whole is well capitalised.

If any insurers consider the standard formula not to be

representative of their risk profile, they can apply for USP, a
partial internal model or a full internal model.

58

SCR.

General SCR calculation


causal relationships

We are concerned about
the requirement to allow in
several SCR modules for the dynamic impacts of causal
relationships between risk drivers. Although the
requirement is justified, companies have no experience in
setting such assumptions and this does not form part of
companies’
normal assumption setting policies or toolkit.
Immediate past experience cannot be referred to for
instance and significant judgement will be required. This
could potentially lead to divergent levels of prudence in the
valuation of different companies. Giv
en the extent of expert
judgement required some sort of review and governance
around this application of significant expert judgement
appears to be warranted. Do the FSB or SAM working
groups have any advice which could be given with respect
to the setting

of these (material) assumptions?

See comment 175


59

SCR.1

SCR Structure

The box for Revision Risk should be for Retrenchment Risk

Noted

To be corrected.

60

SCR.1.1.1 (b) pg 6 (see
point above)

“...the best estimate of the corresponding obligations” ,

suggest changing this to “...the best estimate liability in
respect of the corresponding obligations...”

Noted

None


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
25

of
176

61

SCR.1.1.3

(which is in SCR.7.5) p20
Part 4

Assessment of homogeneity

This section reads: “Homogenous groups of policies for
assessing
lapse risk should be defined at a granular enough
level such that no further split of policies within the
homogenous group as a result of a specific product feature
(e.g. ability to review premiums, premium paying pattern,
presence of guarantee, etc.) will

result in a significant
change in the assessment of lapse risk assessed using the
standard formula.”

To comply with this requirement would require companies
to perform the lapse risk assessment on a per policy basis
to check that there is no feature that
significantly changes
the assessment of lapse risk. This defeats the concept of
using homogenous groups and would be practically difficult
to implement. This statement should be altered to say that
product features that may have an impact on lapse risk
sho
uld be considered when determining the appropriate
homogeneous groups.

Agreed.

Add the following to the
1st bullet under
SCR.1.1.3 (which is in
SCR.7.5):

“… This does not imply a
re
-
calculation per policy,
but rather, taking
proportionality into
account, i
dentifying
those features that
distinguish groups that
are known to behave
differently under
stressed circumstances
or where the impacts of
similar behaviour are
very different.”


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
26

of
176

62

SCR.1.4.1


Definition of own funds is ambiguous, should it include or
exclude the subordinated liabilities?


Replace par. SCR.1.4.1. with the
following two paragraphs.

For several sub
-
modules
the calculation of the
capital requirement is
scenario
-
based: The
capital requirement is
determined as the
impact of a specified
scena
rio on the level of
Basic Own Funds (BOF).

The level of Basic Own
Funds is defined as the
difference between
assets and liabilities. As
explained above, the
liabilities may be
calculated without
including the risk margin
of technical provisions.
Furthermor
e, the
liabilities should include
subordinated liabilities
and BOF should be
reduced with any
exclusion from Own
Funds. The change of
BOF resulting from the
scenario is referred to as
be positive where the
scenario results in a los
s
of BOF.


63

SCR.1.4.1

SCR Structure: Scenario
The definition of basic own funds in SCR 1.4.1 and in section
OF3.1 excludes participations in financial and credit
Basic Own Funds were adjusted in
the QIS2 templ
ate to take into
account participations in financial
and credit institutions (refer to row
No change required
.


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
27

of
176

based calculations

institutions, as well as other exclusions. However, the
templates that
were used for QIS2 included participations
in financial and credit institutions in basic own funds. We
recommend that the templates be changed so as to be
consistent with the definitions in the specification.

171 of the “Balance Sheet” tab in the
template).

64

SCR.1.4.2 c

Part 2


We assume that the requirement to zeroise

a negative
capital requirement is at overall insurer level rather than
product level, but this should be clarified in the text.

Agreed

Clarification to be added

65

SCR.1.4.4 a

Refers to SCR 2.1C, which does not appear to exist.

Noted. This paragraph
refers to the
“Standardised Approach” under
SA

QIS2 which is not used in SA

QIS3.

Reference to be
removed.

66

SCR.10

Operationally ring
-
fenced funds where non
-
discretionary
participation business is included and their profits allocated
to discretionary
participation funds


Allowance should be made in the specification and QIS
spreadsheet to appropriately allow for operationally ring
-
fenced funds consisting of both discretionary participation
and other business, and where profits/losses of the other
busin
ess are allocated to the discretionary participation
business within that fund.



Agreed. Our interpretation of these
operationally ring fenced funds
would be that the non
-
discretionary
business should be valued as part of
that ring
-
fenced fund.

None


67

SCR.10.

The proposals made are more reflective of legal
-
ring



Composite

comments
for SA QIS3 Draft Technical Specifications


Page
28

of
176

RFF

Part 2 of 6

General

Page 34


fencing than that of the current situation of operational ring
fencing. The Promoter still have the option of using funds
from other cells to finance short
-
coming

in cells that are not
financially sound. This is a very important difference and
should be considered.

This is not our understanding of how
cell arrangements work on an on
-
going basis.

None required.

68

SCR.10.1.1

States that Approach B from QIS2 will
apply, but on page 43
Approach B has been deleted.

There is a large overlap between
Approach A and Approach B. Due to
the layout of the specification, it was
easier to adjust the wording in
Approach A to reflect Approach B.
This was due to the wording used

in
SA QIS2 for Approach B mainly just
refers back to Approach A

No change required

69

SCR.10.1.1

Ring fencing

With
-
profit funds should not be treated as ring
-
fenced
funds, unless there are actual legal or operational
restrictions on the backing capital.


Agreed. However, our understanding
is that most with
-
profit funds are
operationally ring
-
fenced, as there
are restrictions on how the assets in
these funds or cash flows from these
funds can be allocated.

Nevertheless, this is an area that
needs further d
iscussion to ensure
that the correct treatment is
obtained.

None

70

SCR.10.1.3

Part 2

We would like to understand the context of and motivation
for the requirement to produce in addition the SCR on the
basis of an internal model (whether

in the model approval

The intention is that the provision of
such information is voluntary.


Further clarity in the
technical specification to

Composite

comments
for SA QIS3 Draft Technical Specifications


Page
29

of
176

Ring
-
fenced funds

process or not). Is this compulsory or voluntary? This
requirement is not detailed anywhere else in the QIS 3
technical specification. Is the internal model SCR only
required in respect of ring
-
fenced funds, or for the total
business
? How will this information be used?

be provided.


71

SCR.10.1.4


The general principle of ring
-
fenced funds set out here
applies to
many insurance contracts in the industry where
profits have a limited ability to absorb losses within other
parts of the business.


An example of this would be reinsurance treaties with
profit share arrangements particular to a specific treaty. So,
profit
made on a treaty is shared with a specific insurer
based on its own specific experience.


We would argue that such contracts with specified profit
sharing mechanisms be considered ring fenced funds
because of their limited ability to absorb potential losse
s in
other parts of the business.



SCR 10.2.14 is not specifically aimed
at ring
-
fenced funds other than cells
and discretionary participation
funds. Where insurers deem such
ring
-
fencing to exist, they should
apply it.


None proposed.

72

SCR.10.1.4

Reference to SCR4.2 is incorrect


should be pointing to
another section.

Reference to be amended.

Reference to be
amended.

73

SCR.10.1.4

The last paragraph refers to section SCR.4.2. We suppose
the section should be SCR.10.2.

Reference to be amended.

Reference to be
amended.


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
30

of
176

74

SCR.10.2.10

SCR.10

Part 2

Ring Fenced Funds


It is our opinion that South African style discretionary
participating business should not be included in the
definition of ring
-
fenced funds. The defining characteristic
of a
ring
-
fenced fund is a restriction on assets and liabilities
which would lead to restricted own funds. Any
undistributed investment return that will be used to
support future bonus declarations is fully included in the
technical provisions. No part of own f
unds needs to be ring
-
fenced for this purpose. These products are similar and
should be treated the same as linked business where policy
benefits are determined by a specified pool of assets, i.e.
operational ring
-
fencing.

Operational ring
-
fencing does not

restrict the diversification
of shock impacts on future profits. It may only restrict the
extent to which assets in a ring
-
fenced fund may be utilised
to absorb losses in another fund.

An alternative approach, if SA style discretionary
participating busin
ess is regarded as ring
-
fenced, which was
previously suggested by the FSB, was to conclude that the
ring
-
fenced or restricted Own Funds for such discretionary
participating business is zero. If this approach is still
acceptable, then it would be helpful to

state this in QIS3.



See response to comment 69


None

75

SCR.10.2.10 & SCR
10.2.13

Part 2

Both refer to SCR 4.3. The reference does not look correct.


References still to be updated.


References to be
updated.


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
31

of
176

76

SCR.10.2.12

Ring
-
fencing and
discretionary participation business


This paragraph seems to imply that all discretionary
participation funds should be viewed as a ring
-
fenced fund,
similar to the requirements for the approach used to test
ring
-
fencing in QIS 2. We believe that the spec
ification
should be updated to specifically exclude all discretionary
participation business which are not viewed as
operationally or similarly ring
-
fenced




See response to comment 74



See response to
comment 74.

77

SCR.10.2.5

Page 35

RFF

The
definitions are not completely correct. Revised
definitions will be provided by the 1st parties working
group.


Noted

Definitions to be
updated

78

SCR.10.2.6 (a)

Incomplete sentence, the word “risk” is missing

Noted.

Cosmetic changes

79

SCR.10.2.6 (a)

Sentence incomplete

Agreed.

Sentence to be
completed

80

SCR.10.2.6 a

RFF

Part 2 of 6

Page 34

The sentence is incomplete.


Noted.


Cosmetic changes.


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
32

of
176

81

SCR.10.2.6(a)

Sentence is incomplete

Agreed.

Sentence to be
completed

82

SCR.10.2.7


We welcome the
proposed change to the QIS. However we
do note that:




There is no guidance as to how this figure should
be calculated which could lead to inconsistency within the
industry;



Ancillary capital should also be allowed for in the
case of 1
st

party cells; particularly where no policy limit
exists or the policy limit exceeds net assets within the cell.



Further guidance to be provided as
to how to recognise the ancillary
own fund item.


Guidance to be updated.

83

SCR.10.2.7

Calls by cells for

additional funding from owner


It is specified that conditions where a cell owner can be
called upon for additional funding can be recognised as
ancillary own funds. However, this would most likely be
restricted in use to those specific cells. Guidance is

needed
on how to deal with this restricted nature.



The intention is to limit the
recognition of such items to
individual cells, and not consider its
transferability. This to be further
explained in the technical
specifications.

Also, note that due to
the nature of
the calculation, any excess own
funds over the SCR in a cell may not
be carried over to the promoters
own funds.


Update for clarity on the
recognition of ancilliary
own funds.

84

SCR.10.2.7

RFF


The introduction of the
use of shareholder guarantees is

The ancillary o
wn funds will only
qualify as tier 2 or tier 3. These tiers

Further clarity to be

Composite

comments
for SA QIS3 Draft Technical Specifications


Page
33

of
176

Part 2 of 6

Page 34

welcomed. We however ask for more clarity on the Rand
amount allowed as it could create a situation where
excessive amounts are introduced that is not realistic; or is
this subject to the insurer to decide.

are restricted in terms of a maximum
% of SCR.

Also, note that due to the nature of
the calculation, any excess own
funds over the SCR in a cell may not
be carried over to the promoters
own funds.

provided.

85

SCR.10.2.8

Reference is made to section SCR4.3. We suppose the
section should be SCR.10.2.

Reference to be amended.

Reference to be
amended.

86

SCR.10.3


The approach selected for QIS 3 was contained in the
CEIOPS
position papers.


We would like to note that this approach is tailored towards
life insurance and is appropriate where there are a few ring
fenced funds within a life license.


The approach did not specifically consider its impact where
a large number of r
ing
-
fenced funds are in operation within
a non
-
life license, particularly for our cell captive business.


From a practical perspective, large
number of arrangements are catered
for by allowing a proxy method for
estimating SCR’s per cell. Further
impact an
d alternative methods to
be considered post
-
SA QIS3.


No change required.

87

SCR.10.3

General comment


capital requirement for ring
-
fenced
funds as it stands will significantly affect the 3rd party cell
captive market. Solvency ratios are likely to
decrease
substantially, especially in the case where a large number of
ring
-
fenced funds are present.

Comment is noted. The SCR
calculation for ring
-
fenced funds is
still evolving. Information from SA
QIS3 will help inform the evolution

No change required


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
34

of
176

Diversification between ring
-
fenced funds is likely
significantly more than that obtained via various lines of
business on a single licence


this is not
fully catered for
within the current proposed capital calculation.

88

SCR.10.3.2 (f)

Second sentence, word “is” should be “if”. Also, the
sentence starting with “The diversified total SCR….” should
be moved to become the second last sentence in the
paragraph because the helper spreadsheet will assist in
calculating the undiversified SCR, n
ot the diversified SCR as
this will be calculated in the standard formula anyway.


Noted.


Cosmetic changes.

89

SCR.10.3.2 f


Whilst we acknowledge that the regulator wishes to
recognise that the capital within the cells is not easily
transferable, the
promoter capital is transferrable.


Where there are a large number of cells involved, targeting
a 1 in 200 year capital requirement using a non
-
life
calculation, summing this total and comparing the
diversified capital requirement would yield a significant

diversification benefit. Depending on the size and number
of cells, it may not be possible to have promoter capital to
cover this notional diversification benefit.


Without recognising the diversification benefit that exists
between the various cells, the

promoter capital
requirement (to fund any deficits) is overstated.


The transferabili
ty of capital within
the promoter cell is acknowledged in
the current method. Further impact
and alternative methods to be
considered post
-
SA QIS3.

None



Composite

comments
for SA QIS3 Draft Technical Specifications


Page
35

of
176


In accordance with SCR 10.2.6, we propose that diversified
capital at the license level is measured and allocated to
every ring
-
fenced cell using an appropriate proxy.


90

SCR.10.3.2 f

RFF

Part 2 of 6

Page 38

The sentence starting with “The diversified total SCR….”
should be moved to become the second last sentence in the
paragraph because the helper spread sheet will assist in
calculating the undiversified SCR.


Noted.


Cosmetic changes.

91

SCR.10.3.2 f

Type: Line 4, sentence should read “This is the total SCR for
the insurer if no diversification…”

Agreed

Change to be made in
line with comment
received.

92

SCR.10.3.3

Definition of SCRNRF


the word “rig” should be “ring”


Noted


Cosmetic changes.

93

SCR.10.3.3

RFF

Part 2 of 6

Page 39


Two problems with the
approach presented:

1.

The way the formula is presented makes the
total insurer SCR become dependent on the amount of own
funds outside ring
-
fenced funds. The higher the own funds,
the lower the SCR. This then becomes a circular reference,
i.e. if the
promoter injects further capital, the SCR reduces.
Ideally, the total SCR should not be driven by the amount of
own funds.

2.

If the amount of own funds outside the ring
-
fenced funds is lower than the SCR outside the ring
-
fenced
funds, the total SCR for the i
nsurer would be higher than

See response to comment 89.



Composite

comments
for SA QIS3 Draft Technical Specifications


Page
36

of
176

the undiversified SCR. Ideally, the maximum SCR should be
the undiversified SCR.

Suggested way forward:

To address the above two problems, we suggest that the
total insurer SCR should be the diversified SCR plus the sum
of all t
he shortfalls resulting from the ring
-
fenced funds.
This approach ensures that the promoter puts up any
capital shortfall under the ring
-
fenced funds.

This suggested approach would encourage insurers to
ensure that all their cells are fully capitalised, ot
herwise the
promoter would be required to top
-
up any shortfalls.

94

SCR.10.3.3

Two problems with the approach presented:


1.

The way the formula was presented in QIS2
template makes the total insurer SCR become dependent
on the
amount of own funds outside ring
-
fenced funds. The
higher the own funds, the lower the SCR. This then
becomes a circular reference, i.e. if the promoter injects
further capital, the SCR reduces. Ideally, the total SCR
should not be driven by the amount of
available own funds.


2.

If the amount of own funds outside the ring
-
fenced funds is lower than the SCR outside the ring
-
fenced
funds, the total SCR for the insurer would be higher than
the undiversified SCR. Ideally, the maximum SCR should be
the undiversifi
ed SCR.


See comment 89.


No change requi
red.


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
37

of
176


Suggested way forward:

To address the above two problems, we suggest that the
total insurer SCR should be the diversified SCR plus the sum
of all the shortfalls resulting from the ring
-
fenced funds.
This approach ensures that the promoter puts up
any
capital shortfall under the ring
-
fenced funds.

This suggested approach would encourage insurers to
ensure that all their cells are fully capitalised, otherwise the
promoter would be required to top
-
up any shortfalls.



95

SCR.10.3.4

Contains the word “oben”


Noted.


Cosmetic changes.

96

SCR.10.3.5 b


While the theory underlying the removal of own funds in a
ring fenced fund from the overall funds seems sound, the
practical consequences for a large cell captive is
as follows:




The proposed SCR calculation relies almost
entirely on the capital strength of the promoter cell.



The SCR cover for a cell captive insurer will
always be extremely low, regardless of how well the
individual cells are capitalised.



The SCR cover

will be misleading (and

This method is applied in this way to
reco
gnise that although some cells
may be extremely well capitalised,
that doesn’t provide any comfort to
those cells which are not. The SCR
cover done in this way considers
capital which is free from
restrictions, similar to that of other
insurers. Further im
pact and
alternative methods to be
considered post
-
SA QIS3.


No change required.


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
38

of
176

understated) when individual cells are well capitalised.



SCR coverage ratios of cell captives will need to
be considered separately from SCR coverage ratios of other
short
-
term insurers.


97

SCR.11.2.6

SCR 11.2.6 should be moved to IRM.

The IRM section already has a similar
paragraph. SCR 11.2.6 should rather
replace “insurance risk mitigation
techniques” with

“financial risk
mitigation techniques”

Replace “insurance risk
mitigation techniques”
with “financial risk
mitigation techniques”
in SCR 11.2.6

98

SCR.11.2.6

Should this statement be part of insurance mitigation, SCR
12?

See reply to comment no. 97

See
reply to comment
no. 97

99

SCR.12

IRM

Part 2 of 6

General

Page 52

Guidance is not provided how collateral held for insurance
risks should be treated.

Noted

Guidance to be provided
in the guidance manual.

100

SCR.12.5.1

IRM

Part 2 of 6

Page 52

E, F and G
appear not to be part of the ….”meet one of the
following requirements” and should be corrected.

Refer to comment 102. above.

Refer to comment 102.
above.

101

SCR.12.5.1 f

IRM

Part 2 of 6

Page 52

Uncertainty exists if national or international ratings
should
be used.

Please refer to comment 149.

Please refer to comment
149.


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
39

of
176

102

SCR.12.5.1.

Part2


Not all items listed in SCR12.5.1 are necessary
requirements, may need to re
-
word the starting sentence
of this point.


Noted.

SCR.12.5.1 to be
reworded as
follows:

“For the purposes of
SA

QIS3, providers of
insurance risk mitigation
should be an insurance
or reinsurance company
regulated under SAM or
an insurance or
reinsurance undertaking
not regulated under
SAM but where the
reinsurance has been
approved b
y the South
African regulator.”


and


Items (c)
-
(g) to be placed
under a separate
paragraph as (a)


(e)
with leading sentence:

“The following specific
requirements must be
met by providers of
insurance risk mitigation
contemplated in
SCR.12.5.1 above:“

103

SCR.13.10


The current proposed formulae don’t consider the policy
Agreed. Policy limi
t should be taken
into account when determining the
Clarification to be added


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
40

of
176

limit. The policy limit can often substantially reduce the
exposure of the insurer. The use of a net retention is not
appropriate in the presence of a policy limit.


Net Aggregate Retention

104

SCR.13.12, SCR.13.14
and SCR.13.17

The non
-
life underwriting risk SCR simplification formula
presented does not allow for any diversification. We
acknowledge that the lines of business within one first
party structure may not be independent and therefore the
undiversified calculation within a single first party structure
is supported (a correlation matrix could be considered here
but for simplicity
this is not recommended).


Individual first party structures are however independent of
each other, and in this regard we recommend that some
form of diversification should be allowed between the
individual first party structures. This can be achieved by
c
alculating the square root of the sum of squares for
individual first party structures’ non
-
life SCR. We
recommend the formula below.


Furthermore, we do not see the reason for introducing the
requirement that the non
-
life SCR for each first party
structur
e should be at least 80% of the Liability class net
aggregate retention. We recommend that this requirement
should be removed. If there is a compelling reason to
No changes from SA QIS2
have been
made yet to the First Party Insurance
Structures.
Comments

will be taken
into account by the Working Group

New formula included in
simplification


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
41

of
176

support keeping this requirement, then we recommend
that this should be applied after the tota
l non
-
life
underwriting risk SCR has been calculated for all first party
structures, not individually per first party structure.


Suggested way forward:

Using the Inputs defined in section SCR.13.10, we
recommend the following formulae to replace the
calculation formulae under section SCR.13.12:











This revised formulae will ensure that those insurers with

Composite

comments
for SA QIS3 Draft Technical Specifications


Page
42

of
176

multiple first party structures will have a diversification
benefit between (but not within) the said structures.

105

SCR.13.17

If the above suggested approach is accepted, this section
should be rephrased as
follows:

“The simplification does not allow for diversification
between lines of business within a first party insurance
structure but diversification is allowed between different
first party insurance structures within one legal entity”.

Diversification a
llowed for some
arrangements

Specifications have been
updated

106

SCR.13.2.1


A worked example to demonstrate why the need for
inconsistency in premium recognition between life and
short
-
term insurers is requested. This should compare a:



Short
-
term
insurance policy that has an annual
contract boundary and premiums becoming due
monthly; with



A 5 year term assurance product that has
premiums becoming due monthly.


Either confirm the intention of the phrase “...or expected to
become due...” or delete it

from the requirement.



Section SCR.13.2.1 could not be
found


it is unclear what the
comment relates to

None

107

SCR.13.7

This section is not relevant (as it refers to QIS1) and can
therefore be deleted.

Even though relating to SA QIS1,
paragraph 13.7
provides useful
background information. Paragraph
13.8 has been added to provide
None


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
43

of
176

feedback on SA QIS2 exercise

108

SCR.13.8

The simplification will be tested as a 2nd attempt, not first
attempt as stated.

Noted

To be corrected.

109

SCR.14.2.1

Participations

The equity shocks detailed in 5 are incorrect


should be
44%, 52% and 58% instead of 37%, 47% and 53%
respectively. What is currently shown are the SA QIS2
shocks, so they have not been updated for SA QIS3.

Noted

Specification to be
updated

in line with
market risk section.


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
44

of
176

110

SCR.14.3.1

Part2


The additional information is requested at 31 December
2012, 2011 and 2010; if a QIS is completed at a different
date (e.g. June) can the additional information be supplied
at that date instead.


Noted.

To be reworded as:

“(a) The value according
to subsection SCR.14.2
as at the reporting date
chosen by the insurer for
SA QIS3.”


Composite

comments
for SA QIS3 Draft Technical Specifications


Page
45

of
176

111

SCR.2

Covering letter: Strategic
Participations


Strategic participations:

o

We believe that it is reasonable for
strategic
participations in the same industry to receive the full equity
shock and no diversification benefit (because they might be
in stress at the same time as the insurer). However, we
have a number of strategic participations which are not in
the same

industry as us and which would not necessarily be
stressed at the same time, which we think should be
treated as other equity under the equity risk sub
-
module
and therefore receive diversification benefits.

o

We think we should be allowed to include certain

participations in the equity risk sub
-
module.

o

We also believe that a look through approach will
be very complex, difficult and time
-
consuming with little
potential benefit, in our case.



Agree that participations which are
not insurance related should be

included with equity stresses, and
should hence receive diversification
benefit. For this purpose,
underwriting management agents
and brokers should be deemed to be
insurance related

Specification to be
updated

112

SCR.2


The lay
-
out of the section is
confusion. We propose the
following changes:


SCR.2.1 Strategic participations are those participations that
the insurer intends to keep for a long period of time.

SCR.2.2 Strategic participations in financial and credit
institutions (which include banks a