how will the ifrs 17 discount rate impact the balance sheet ......7 ifrs 17 bottom-up discount rate...
TRANSCRIPT
20/11/2018
1
How will the IFRS 17 discount rate impact the balance sheet and profitability for insurers?Dylan Grainger, Prudential plcMatteo Ricciarelli, PwC
20 November 2018
Agenda
• IFRS 17 background and discount rate requirements
• Top-down: what is the reference portfolio and how can it change?
• Bottom-up: what is a liquidity premium and how can it be applied to liabilities?
• IFRS 17 discount rates compared to Solvency II
• Discount rate driven accounting mismatches under IFRS 17
20 November 2018 2
20/11/2018
2
IFRS 17 background and discount rate requirements
20 November 2018 3
IFRS 17 – Brief overview
• IFRS 17 was published in May 2017, 20 years after the IASB project on insurance contracts was initially announced
• Replaces interim standard IFRS 4 effective 1 January 2021 (?)
• Another change to insurance companies following the implementation of Solvency II
20 November 2018 4
20/11/2018
3
IFRS 17 – Balance Sheet
20 November 2018 5
Current unbiased probability weighted
estimate of future cash flows
Contractual Service Margin
Risk Adjustment
Fu
lfill
me
nt
ca
sh f
low
s
Time value of money
Unearned profits recognised in income statement over coverage period (measured at
locked-in economics for general model & current economics for variable fee approach)
Reflects compensation for uncertainty from non-financial risk
Discounting future cash flows using “Top-down” or “Bottom-up” rate that takes account
of illiquidity
Explicit, unbiased and probability weighted estimates
What does IFRS 17 say about discount rates?
20 November 2018 6
36Discount rate should be consistent with:• the timing, currency and liquidity of the liabilities• observable current market prices (where available)
B80
Discount rate may be calculated using a “Bottom-up” approach (start with risk free curve and add a liquidity premium)
B81Discount rate may be calculated using a “Top-down” approach (start with the total yield on a reference portfolio and deduct credit risk)
The standard gives a choice of two methodologies to set the discount rate:
20/11/2018
4
Top-down discounting
20 November 2018 7
IFRS 17 Top-down discount rate
20 November 2018 8
Liquid risk free curve
Equity risk premium
Liquidity premium
Liquid risk free curve
Liquid risk free curve
Expected default
Unexpected default
Mismatch adjustment
Implied liquidity premium
Tota
l yie
ld o
n “
refe
renc
e” p
ort
folio
Equity backed Debt backed
“Top-down”
• Starting point is a reference portfolio of assets
‒ “IFRS 17 does not specify restrictions on the reference portfolio of assets” which can lead to very different outcomes as
‒ “not required to adjust yield curve for differences in liquidity characteristics” between liabilities and reference portfolio
• September TRG clarified that reference portfolio can be actual asset portfolio over time, but could use an appropriate ‘hypothetical portfolio’
• Deductions on asset yields are also required for other factors not relevant to the liabilities
= Discount rate
20/11/2018
5
Deriving credit risk deductions (CRD)
• Disentangling the incidence of credit risk from credit spreads is generally a difficult task and IFRS 17 does not prescribe a method to do so
• Academic literature illustrates that structural models tend to underestimate credit spreads and term this outcome as the “credit spread puzzle”
• Credit risk is a risk factor that influences the credit spread while the credit spread is made up by systematic and idiosyncratic risk factors, and these risk factors have different explanatory power:
– Liquidity
– Business cycle
– Monetary policy
20 November 2018 9
Potential approaches to CRD
• In keeping with its principle based nature, the standard does not lay out a methodology to derive CRD
• Different potential approaches include:
– Historical defaults
– Distribution-based derivation of expected/unexpected losses
– Credit Default Swap
– EIOPA’s fundamental spread
– Fair value spreads derived from structural models á la Merton (1974)
20 November 2018 10
20/11/2018
6
Mismatch risk
• The discount rate should reflect the characteristics of the liability. Practically, both Top-down and Bottom-up approaches start from asset considerations
• The implicit assumption is that assets appropriately reflect the features of the liability product in terms of cash flow timing, currency and liquidity
• The mismatch risk adjustments refer to those circumstances where this assumption doesn’t hold e.g. cash flow mismatches between asset portfolio used to derive yields and liability
20 November 2018 11
Bottom-up discounting
20 November 2018 12
20/11/2018
7
IFRS 17 Bottom-up discount rate
• Starting point is a liquid risk-free yield curve.
‒ Not prescribed. Could use EIOPA risk free curves?
‒ What is the ultimate forward rate?
• Calculate a ‘market price for illiquidity’ from suitable market instruments to get 100% liquidity premium
• How to balance using market observable data with assessing ‘market price of illiquidity’ which isn’t readily observable?
• Analysis on the predictability of cashflows is key to support the argument for the use of and size of liquidity premium
20 November 2018 13
Liquid risk free curve
Liquidity premium
“Bottom-up”
= Discount rate
Liquidity premium: background
• Last significant industry efforts on illiquidity premium were in 2009 as part of QIS5
• Matching Adjustment and Volatility Adjustment were adopted into Solvency II, and illiquidity premium became a less urgent topic
• QIS5 used a combination of three approaches:
‒ Covered Bonds
‒ CDS Negative Basis
‒ Structural Debt Models
• There are other alternative approaches in financial literature based on bid-ask spreads and pricing a hypothetical liquidity swap, but these are less established and more conceptual
20 November 2018 14
20/11/2018
8
Covered Bonds
20 November 2018 15
Source: ICE Index Platform
-20
-10
0
10
20
30
40
50
60
70
80
Ba
sis
Po
ints
$ Liquidity Premium (US Large Cap Covered Bond OAS) £ Liquidity Premium (Sterling Non-Gilt Covered OAS) € Liquidity Premium (Euro Covered Bonds OAS)
CDS Negative Basis
20 November 2018 16
-50
0
50
100
150
200
250
300
Sep
201
1
Jan
20
12
Ma
y 20
12
Sep
201
2
Jan
20
13
Ma
y 20
13
Sep
201
3
Jan
20
14
Ma
y 20
14
Sep
201
4
Jan
20
15
Ma
y 20
15
Sep
201
5
Jan
20
16
Ma
y 20
16
Sep
201
6
Jan
20
17
Ma
y 20
17
Sep
201
7
Illiquidity Premium EUR CORP CDS 5yr
EUR Corp Bond OAS (1-10 yr)
-
50
100
150
200
250
Se
p 2
011
Jan
20
12
Ma
y 2
012
Se
p 2
012
Jan
20
13
Ma
y 2
013
Se
p 2
013
Jan
20
14
Ma
y 2
014
Se
p 2
014
Jan
20
15
Ma
y 2
015
Se
p 2
015
Jan
20
16
Ma
y 2
016
Se
p 2
016
Jan
20
17
Ma
y 2
017
Se
p 2
017
Illiquidity Premium US CORP CDS 5Y
US Corp Bond OAS (1-10 yr)
EUR USD
Source: ICE Index Platform, Bloomberg
20/11/2018
9
Structural Model
20 November 2018 17
-
50
100
150
200
250
Illiquidity Premium Structural Model Fair Value Spread iBoxx Index OAS
GBP
IFRS17 – Bottom-up discount rates
• No clear preferred method
• QIS 5 approach was based on a ‘Simplified Formula’ based on a combination of these approaches:
• Hard-coded values in the QIS 5 approach were the results of a calibration to a bond index using EUR, GBP and USD data over 2005-2009. Those values may require updating.
• The formula produces a flat liquidity premium to add to the risk free rate term structure
20 November 2018 18
, % . %
20/11/2018
10
A term structure of illiquidity premiums
20 November 2018 19
0
37
5869 72
89105
0
20
40
60
80
100
120
0 5 10 15 20 25 30 35
bp
s
Years
Illiquidity Term Structure (QIS 5) 5% 20%
55%
20%
Cash
EEA Sovereign
Corporate Investment Grade
Illiquids
• For example a term structure of illiquidity premiums has been obtained by applying the QIS 5 formula to the spread of different maturity buckets of this allocation
How to apply liquidity premium to liabilities?
• Qualitative assessment e.g. similar to QIS5 bucketing approach?
… Or is quantitative testing required to assess the predictability of the liability cash flow profile e.g. Compare base to stressed liabilities under a range of scenarios
20 November 2018 20
100%LQP
75%LQP
50%LQP
e.g. Annuities e.g. Unit linked protection e.g. With-profits
20/11/2018
11
Top-down vs. Bottom-up: illustrative comparison
20 November 2018 21
Illustrative example
20 November 2018 22
• The portfolio below resembles the high-level allocation of a Matching Adjustment Fund and would be considered as an adequate starting portfolio to derive a top-down discounting
Asset Class Allocation Spread over Swaps
Cash 5% -
Gilts 20% 1.4%
Corporate Bonds 55% 3.2%
Illiquid Assets 20% 4.0%
Total 100% 1.59%
20/11/2018
12
Illustrative example cont.
20 November 2018 23
159
60
114 115
21
020406080
100120140160180
GrossPortfolioSpread
IlliquidityPremium:
Bottom - Up(QIS 5)
IlliquidityPremium:
Top - Down
MatchingAdjustment
VolatilityAdjuster
IFRS 17 Solvency II
Top-down versus bottom-up?
Bottom-Up• Volatility primarily driven by the volatility of the spread of the
reference portfolio/index used to derive the illiquidity premium.
• Wide host of methodologies available introduces also some model risk, i.e. different answers may result from different methods to calculate the illiquidity premium
• In case representative market indices are used to proxy holdings, then the approach is less sensitive to ALM inputs (e.g. SAA reviews). However, some basis risk may arise when the reference portfolio/index does not mirror the actual asset holdings
• Likely to be conducive to a greater alignment with Solvency II discounting
• Forward-looking choice of reference discount curve will minimise the issues stemming from Libor reform. However, transitional calculations still would have to refer to Libor
Top-Down• The volatility is predominantly driven by the actual yield of the
reference portfolio and, in particular, its spread over the reference rate.
• Spreads may change due to systematic and/or idiosyncratic risk of default and downgrades of the securities held in the underlying portfolio as well as its asset allocation.
• Expected losses (i.e. default probabilities and recovery rates) are updated infrequently and are based on long-term experience. Therefore, it may have lesser impact in terms of volatility of the discount
• Tactical ALM choices (e.g. a duration mismatch) may trigger an extra-deduction, in addiction to the one for expected/unexpected losses.
20 November 2018 24
20/11/2018
13
Discount rate driven accounting mismatches under IFRS 17
20 November 2018 25
Accounting mismatch 1 (general model)
20 November 2018 26
Negative FCF (asset)
CSM
Negative FCF (asset) CSM
Interest rate drop by end of year 1Day-1
• Accounting mismatch introduced as interest rate fall increases FCF asset, but CSM is not impacted by change in interest rates as measured based on locked-in discount rate. This causes P&L and equity volatility.
• P&L volatility could potentially be minimised by electing the OCI option. The effectiveness of the OCI option should be explored alongside IFRS 9
20/11/2018
14
Accounting mismatch 2 (general model)
20 November 2018 27
• The use of a locked in discount rate for the CSM in the general model means that the impact of operating assumption updates is absorbed in the CSM at the locked-in rate.
• As the BEL is measured at the current rate the difference between the assumption change measured at the locked-in rate and current rate is reflected in the P&L and can distort the current period result.
CSM
FCF
Before assumption change Impact of assumption change
FCF
CSM
at current rate (e.g. 3%) at locked-in rate
(e.g. 6%)
After assumption change
Impacts P&L but not through insurance service result!
OCI: Income statement implications and IFRS 9
• Under the general model, changes in the liability due to changes in the discount rate flow through P&L or are disaggregated between P&L and OCI (“OCI-option”)
• Under the OCI option, the insurance finance income & expenses is stable over time:
– Unwind of time value of money: calculated by using the curve at initial recognition
– Interest accretion on the CSM: calculated by using the curve at initial recognition
• Different classification of assets based on the new IFRS 9 standard may also result in different volatility profiles for P&L and OCI
20 November 2018 28
20/11/2018
15
Conclusions
20 November 2018 29
Conclusions
• IFRS 17 is a principles based standard with explicit choice to use a “top-down” or “bottom-up” approach to setting discount rates
• Top-down: Expected to use actual assets backing liabilities (not a new concept) with main challenge around estimating unexpected default allowance
• Bottom-up: No market consensus on approach to incorporating allowance for liquidity premium but will likely have increasing focus over next year or so
• Accounting mismatches will exist where locked-in discount rates are required
• Analysis of the simultaneous impact of the applications of IFRS 9 &17 on P&L and profitability are required
20 November 2018 30
20/11/2018
16
20 November 2018 31
The views expressed in this presentation are those of invited contributors and not necessarily those of the IFoA. The IFoA do not endorse any of the views stated, nor any claims or representations made in this presentation and accept no responsibility or liability to any person for loss or damage suffered as a consequence of their placing reliance upon any view, claim or representation made in this presentation.
The information and expressions of opinion contained in this publication are not intended to be a comprehensive study, nor to provide actuarial advice or advice of any nature and should not be treated as a substitute for specific advice concerning individual situations. On no account may any part of this presentation be reproduced without the written permission of the authors.
Questions Comments