Lloyd s NED Reserving Forum 2017

Transcription

1 Lloyd s NED Reserving Forum Richard Rodriguez and Catherine Scullion

2 Contents Introduction Reserving oversight at Lloyd s Reserving oversight in MAs Focus Areas Key Takeaways Questions/discussion 2

3 Introduction

4 Catastrophe versus reserve risk Lloyd s Internal Model Central Fund loss 4

5 Lloyd s historical results 1950 to 2016 UW better in recent years Establishment of business planning and new entrance process NZ, Japan EQ, Thai Flood WTC Attacks, reserve deteriorations Hurricane s Katrina, Rita & Wilma Major losses: Hurricane Betsy (1965), 1974 Super Outbreak 148 tornados in one day, Piper Alpha (1988), Hurricane Hugo (1989), the San Francisco Earthquake (1989), Exxon Valdez (1989) North European storms (1987 and 1990), Typhoon Mireille (1991), Hurricane Andrew (1992), Northridge Earthquake (1994), WTC (2001), Hurricanes Charlie, Francis, Ivan (2004), Hurricanes Katrina, Rita, Wilma (2005), New Zealand, Chile Earthquake (2010), New Zealand, Japan Earthquake, Thailand Flood (2011) Source: Lloyd's Annual Reports, Statistics relating to Lloyd's 2001; Lloyd s data for on three year accounting from 2000 onwards on 5 annual accounting basis.

6 Underwriting or reserving? 6

7 Reserving matters Largest: Item on the balance sheet Single cause of historic failures in US P&C industry Cause of Central Fund hits (even post Equitas) Contributor to underwriting results since 2004 Key linkage with capital and pricing Subject to regulator and rating agent scrutiny and gets due attention! Lloyd's Asset Stack Technical provisions FAL Central 7

8 Particular focus at current time: Rise in combined ratio with major claims costs exceeding prior year releases 2015 Combined ratio % 2016 Combined ratio % 94.4% 3.5% (7.9%) 93.9% 9.1% (5.1%) 14.1% 14.0% 26.0% 26.6% 94.4% 90.0% 90.0% 93.9% 97.9% 97.9%¹ 54.3% 53.3% AY excl. Major claims Prior year Major claims reserve movements Calendar year AY excl. Major claims Major claims Prior year reserve movements Calendar year Attritional loss ratio Acquisition expense ratio Admin expense ratio Source: Lloyd's pro forma financial statements, 31 December AY: accident year. ¹The result and combined ratio has been impacted by the change in the Ogden discount rate in February

9 This is not unexpected Major claims marginally above long-term average 14% 9% (5%) Reserve releases not expected to continue at this level in current conditions Attritional experience in line with recent years 27% 53% 93.9% 97.9% 98% AY excl. Major claims Major claims Prior year reserve movements Calendar year 2016 should not be considered an unusual year of performance 9

10 And permeates the market Split of combined ratio and accident year ratio for ongoing syndicates excl. SPAs (2016: 83, 2015: 82) AY < 100 AY > 100 Release Strengthen CR <100 CR >100 10

11 Reserving oversight at Lloyd s

12 Reserves Ultimate Claims Reserves persist for many years 12

13 with significant exposures to long-tailed lines 100% 90% 80% 70% Accident & Health 60% 50% 40% Aviation Casualty FinPro Casualty Other Casualty Treaty Energy 30% Marine Property (D&F) 20% Property Treaty Specialty Other 10% 6 0% Premium 2016 Reserves (OS + IBNR) Casualty classes reserves being longer-tailed are disproportionate to premium volumes 13

14 Probability and considerable associated uncertainty Generally give rise to distributions with more downside than upside 25% % % % % % 0.0 Reserve ( bn) Reminder: A best estimate is expected to be too low 50% of the time 14

15 Wide range of oversight metrics and processes employed at Lloyd s Top down Qualitative: Market level issues and PMD/claims monitoring Quantitative: Market level projections, allocation to syndicate Syndicate level reserve adequacy Quantitative: Quarterly movements, large losses, benchmark levels Qualitative: SAO and AFR reporting; Minimum standards compliance SAO: Statement of Actuarial Opinion; AFR: Actuarial Function Report Bottom Up 15

16 Detailed monitoring of reserve releases (A slide we show to rating agents) Reserving Minimum Standards No anticipation of releases when approving plans, once reserves established margins are evident Expected to be supported by benign experience/held margins Consistent level of held margin supports releases, given favourable experience Monitoring: Monitor releases (supported) and any deteriorations (sufficient) Look at class of business and year of account source Compare with central view and reserve benchmarking Level of release is not expected to be maintained even in the face of continued good experience - market conditions not expected to support to the same degree 16

17 Current focus area timeline Casualty classes in recent years 17

18 Reserving Oversight in Managing Agents

19 Reserving oversight in Managing Agents Statutory and Lloyd s requirements Reserving Minimum Standards Clear framework within which Managing Agents are required to operate Reserve setting as board responsibility Board attests compliance Valuation requirements for Lloyd s Solvency No Lloyd s requirement to hold margins No credit for releases in business planning 19

20 Releases also monitored at Syndicate level according to benchmark reserve strength Releases (only) split by RBI quartile - MRC Estimate 6% Deteriorations (only) split by RBI quartile - MRC Estimate 1% 4% % 24% 51% 64% 31% Benchmark comparisons are used to assess consistency and relativity. A range of philosophies is present across the market. Clarity and consistency is key. 20

21 Role of NEDs (Reserving context) High level perspective Objective assessment Completeness and consistency Information provided Access to actuarial 21

22 If it looks like a duck Results and information provided should: Look reasonable Be internally consistent and consistent with wider market information Follow processes which are consistent over time Highlight limitations and uncertainties Link to pricing and capital Actuarial methods rely on assumptions and have limitations There s no such thing as a stupid question Actuaries aren t as clever as we make out Don t underestimate the importance of the sniff test! 22

23 Focus Areas

24 Focus areas Example MI Planning versus reserving Actuarial sign-off Uncertainty and margins Key takeaways 24

25 MI should be clear and direct attention appropriately Rate adjusted Gross or net How has reinsurance changed? Differences in business mix Placement type, terms and conditions Process changes Case reserving strength Movements over time Drivers of these: Experience/other? Response to these: Appropriately comprehensive 25

26 MI - Actual versus expected m, net of reinsurance, incurred claims Expected movement Actual movement Actual less expected AvE Movement in ultimate (10) (3) Total

27 MI - Actual versus expected m, net of reinsurance, incurred claims Note the basis Expected movement Actual movement Actual less expected AvE Movement in ultimate (10) (3) Total Totals and year of account provide important feedback information 27

28 MI - Actual versus expected m, net of reinsurance, incurred claims Note the basis Expected movement Actual movement Actual less expected AvE Totals and year of account provide important feedback information Is the experience driven by specific factors it is starting to look systematic Have any assumptions been updated in response? Movement in ultimate (10) (3) Total Should other years be updated? i.e. is 2016 better or just going to emerge later? 28

29 History tells us plans (and reserves) are usually wrong 29

30 History tells us plans (and reserves) are usually wrong 9 pts WARNING: The scale is large 30

31 MI Estimates over time How wrong we ve been in the past is a guide to where we should be now 31

32 MI Estimates over time How wrong we ve been in the past is a guide to where we should be now Note: Looks even more like a duck when rate adjusted 32

33 MI Business planning assumptions For years of account early in development actuarial methods need an assumption on expected performance of the business Often based on business planning Particularly material for Casualty lines (where early experience is minimal) Picture 2017 year-end Business plan % % % 2018? Reduction justified in business planning. Class focuses on niche lines to get rate increases Reduction justified in business planning. Class is marginal and subject to significant re-underwriting 33

34 MI Business planning assumptions Business plan One year later Two years later Three years later % 66% 68% 68% % 65% 67% % 66% 2018? Business performance is adverse compared to plan Business mix/focus and underwriting actions are not having the expected effect Credit for these actions should not be taken in reserving assumptions unless there is strong reasoning to support their affect Questions that should be answered: How have business planning assumptions been used in reserving work? How has performance compared to plan? 34

35 MI Business planning assumptions Business plan One year later Two years later Three years later % 66% 68% 68% % 65% 67% % 66% 2018? Business performance is adverse compared to plan This is the progression the reserving exercise needs to explain Business mix/focus and underwriting actions are not having the expected effect Credit for these actions should not be taken in reserving assumptions unless there is strong reasoning to support their affect Questions that should be answered: How have business planning assumptions been used in reserving work? How has performance compared to plan? 35

36 MI Business planning assumptions If a plan has required aspirational/unproven assumptions on change the suitability of these for reserving should be considered separately and in context of latest position of recent years Latest position % % % % 2018 based on plan assumption, not historical outcome. Not realistic given actual performance of other years. 36

37 Actuarial Sign-off A Statement of Actuarial Opinion (SAO) is required on all open Syndicate years of account Among other uses, this provides the level of reserve margin Lloyd s recognises as capital Lloyd s expects this to be visible to the board: Source of independent challenge Incorporate wider market experience More detailed discussion where views diverge Understand drivers of difference Rotation consideration (New for 2017) 37

38 We expect the board to use actuarial sign-off to support its informed, objective challenge Appropriate in level The board should ensure an appropriate level of internal and external actuarial analysis is provided to support their reserve setting responsibility. Visibility If an external actuary has recommended reserves or has provided an independent estimate of the reserves the Board should consider their analysis and have the opportunity to challenge this. Accessibility A good practice is for non-executive directors to meet the external actuary without the executive board members in attendance. 38

39 Uncertainty and margins Recap: Reserving is subject to significant uncertainty Uncertainty should be highlighted in summary and consider specifics of the business Key uncertainties should be outlined and quantified using stress/scenario testing or distributions of reserves For example: if the expected loss ratio of Cyber is increased from 60% to 70% the class reserves increase by 8%. There is limited history for this class making this material assumption highly uncertain If the impact is material, how certain is the assumption? 39

40 Uncertainty and margins The philosophy to holding reserve margins varies across the market Credit can be claimed for these as part of Lloyd s solvency This credit is only appropriate for genuine margin Expectation of consistency in philosophy and levels of prudence change in this should be conscious and with good reason Capital is available to cover UNEXPECTED movements (volatility): Not those that should have been included in estimates Not poor performance which could have been anticipated based on prior performance 40

41 Uncertainty and margins If uncertainty requires reserves these do not represent margin Real world example 1: Silverstein (WTC) - If a reserve is held for a particular large loss which is uncertain this is not necessarily margin a 90% chance of a loss of 0 and a 10% chance of a loss of 500m has an expected value of 50m Real world example 2: The Ogden discount rate review presented a strong likelihood of changing the rate, which should have been incorporated in best estimates Any use of margin to cover ongoing uncertainty in reserves is a change in basis and should be clearly communicated 41

42 Key Takeaways 1. Ensure the MI allows you to ask the right questions (we ve provided some suggestions) 2. How is experience incorporated in estimates and assumptions 3. Are the links to pricing/business planning appropriate? 4. What are the most material assumptions and how uncertain are these 5. Look out for double-counting of reserves in margins 6. Be nice to actuaries but make sure they can explain why they are wrong and what they re doing about it! 42

43 Questions/Comments

44