Lloyd s Signing Actuaries Forum Henry Johnson, Jerome Kirk & Ben Thomas December 5 th 2014 Lloyd s 1
Erratum PPO section In the presentation Henry Johnson warned actuaries to remember the origins of the ASHE index when they used it in models. It has been pointed out, correctly, that the use of ASHE is based on court decisions to index PPOs to ASHE. Henry apologises for his incorrect comment. Lloyd s 2
Agenda Observations on year to date including margins in reserves SAOs including Large Loss Wordings Regulatory developments including future of SAOs Questions Lloyd s 3
Observations on year to date Lloyd s 4
IBNR m IBNR Burn MRC A vs E results do produce a surplus Ex Cat IBNR burn over 2014 YTD 2007 & Prior 2008 2009 2010 2011 2012 2013 Total 2012 & Prior Pure YOA Gross IBNR Ex Cat Expected IBNR Burn (R/H Scale) Actual IBNR Burn (R/H Scale) Source: Lloyd s Early Warning Analysis Lloyd s 5
IBNR m IBNR Burn with a mixed picture by class when all years of account are combined. Ex Cat IBNR burn over 2014 YTD by COB Gross IBNR Ex Cat Expected IBNR Burn (R/H Scale) Actual IBNR Burn (R/H Scale) Source: Lloyd s Early Warning Analysis Lloyd s 6
IBNR m IBNR Burn for example experience on 2013 YOA Ex Cat IBNR burn over 2014 YTD (2013 YOA only) Gross IBNR Ex Cat Expected IBNR Burn (R/H Scale) Actual IBNR Burn (R/H Scale) Source: Lloyd s Early Warning Analysis Lloyd s 7
In November we were asked about cargo experience Cargo ULRs SBF TPD Best Estimate TPD Held 2005 2006 2007 2008 2009 2010 2011 2012 2013 Year of Account Lloyd s 8
which is a little worse and follows the same overall pattern as other classes Cargo A vs E during 2014 A vs E 2008 & Prior 2009 2010 2011 2012 2013 Year of Account Lloyd s 9
Overall view of A vs E Not as good as 2013 experience Overall no significant concerns Watching latest exposures and some classes Like 2013, no Cats observed yet in 2014 But we have seen some gross deteriorations on New Zealand and Costa large losses little impact Net not immediately concerned Lloyd s 10
Themes Lloyd s 11
Themes PPO update Carried out the first PPO exercise during 2013 (as at 2012Q4) Questionnaire to market Supplemented by Exchanging data and SAOs Results sent to market by individual syndicate Update as at 2014Q2 No further information requested from the market directly Exchanging data, so gross only and most UK Motor and some EL excluded but these are not a major component Case reserves only/ppos in payment Lloyd s 12
Previous exercise: PPO Numbers by risk code Lloyd s 13
Spread over YOA Source: Xchanging data Lloyd s 14
Number of insureds by YOA increasing. 25 20 15 10 5 0 1995 1996 1997 1998 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 As at YE2012 1 2 2 2 4 2 10 3 5 10 15 9 9 7 1 0 0 As at YE2013 1 2 2 2 4 3 14 10 9 15 18 12 15 19 4 1 0 As at 2014H1 1 2 2 2 5 3 16 12 12 16 20 15 18 23 10 1 1 Lloyd s 15
but average reserves per insured stable Insured s Average Reserve Per Claim m 2012 YE 2013 YE 2014 H1 Number of Insureds 75 113 138 Total Case Reserves 37 44 56 Average Case Reserve per Insured 0.5 0.4 0.4 Lloyd s 16
PPOs IBNR/IBNER No consistently updated information on PPOs which are not yet known about Either as a PPO, or as any kind of claim Highly uncertain but need to be clear on assumptions Some information obtained from SAOs we have asked for specific comment to be made Shows that assumptions are broadly consistent with what we have seen in the past: ASHE assumptions Real discount rate Lloyd s 17
PPOs Conclusions Number of PPOs and insureds has increased steadily In line with expectations In the YOAs we would expect to see given court settlement delays Continued reduction in XM premium Case reserves per insured consistent Continue to monitor trends but still no concerns Lloyd s 18
Themes Forward View UK Deafness Practitioners report some deteriorations Consider looking at UK EL Inflation the sustained low inflation environment could change Please consider the implications for reserving Cargo Lloyd s 19
Reserve Margins / Treatment of Bad Debt Lloyd s 20
We did see large reserve releases at half year and from the right places Prior year result : 0.76bn 2/3 rd of releases come from syndicates with a reserve benchmark index of 1 or 2 We do think the reserve surplus reduced compared to y/e 2013 Expect scrutiny of large any releases at year end if margin remaining constant Source: Lloyd s Reserve Benchmarking Index Lloyd s 21
Still not there on reserve margins at half year A great improvement on 2013 but there still were a number of queries Most difference related to margin in bad debt More on next slide but same principle applies Margins in GAAP numbers vs best estimate as per signing actuary Basic concept is if you book more margin than the SAO get the signing actuary involved Lloyd s 22
What are the requirements on Bad Debt? IFoA advisory note originally produced in January 2000 Updated in October 2005 You should pay regard to it with justification you can use different assumptions and in fact Solvency II requires firms not to rely automatically on external credit assessments (Level 2 Article 259) If you believe there is prudence in your bad debt estimate then this needs to be explicitly stated and signed off if to be used as a margin for capital Lloyd s 23
Premium Provision In addition to the view on margin in the earned provisions we expect Signing Actuaries to explicitly consider the impact on margins from the unearned provisions where applicable. For example Higher Future Premiums than the Managing Agent view as a result of differing views in ultimate premium. Difference in view of earned premium due to application of alternative earning patterns (especially for reinsurance) between the Signing Actuary and the Managing Agent. There may also be an impact on the future unearned claims from the above. See QMC Instructions (Section 1.10) for further details. Lloyd s 24
SAOs Lloyd s 25
2014 SAO Review Process Overall performance for all producers by year-end Excellent Good 60% 61% 65% 69% 71% 72% 74% 71% 72% 72% Meets Requirements Poor Unacceptable 2005 2006 2007 2008 2009 2010 2011 2012 2013 Scores based on full review Scores based on 'Lite' review Note: 2011 score rebased for comparison purposes Lloyd s 26
2014 SAO Review Process Grading guidance has been tightened in three sections: 1. Methods 2. Assumptions 3. Results Selection Question Section Excellent Good Meets Requirements (3.17) How well does the report communicate the methods used for producing the estimates? Methods - Description of methods Methods & implicit assumptions underlying the methods are described well. An excellent grading indicates a report which stands out for clarity in communicating methods, and that most or all of following are satisfied: - It is clear what methods have been used for claims projections for each year of account and material class of business, and this includes clear description of approaches in the case of departures from the usual chain ladder and BF approaches. - There is a clear description of methods used to net estimates down for reinsurance. - Methods used for projecting large/cat losses are clearly described, so that it is possible to see which methods have been applied to all large losses/cats. - Methods are described well for ULAE, RIBD and Trust Fund estimates. - For ultimate premiums there is mention of the projected values used and how they have been sourced (e.g. from syndicate's own estimates) or projected. Methods are described in an acceptable level of detail - It is clear which methods have been used for claims projections and how they vary across the book, although specific detail may not be available for every year of account, class of business and large loss. - There is clear description of methods used to net estimates down for reinsurance. - Methods are described broadly for ULAE, RIBD and Trust Fund estimates. - For ultimate premiums there is mention of the projected values used and how they have been sourced (e.g. from syndicate's own estimates) or projected. Methods are described briefly in a summarised form and your understanding of them is adequate. Not all of the points required for 'good' are addressed. (3.18) How well does the report communicate the key assumptions made? Methods - Assumptions Key assumptions are readily identified and well described. As a result, you have a detailed understanding of the key assumptions made. An excellent grading indicates a report which stands out for the clarity with which assumptions are communicated, and that most or all of the following are satisfied: - For claims projections any IELRs or assumed ULRs are provided at a class of business and year of account level, there is a description of any benchmark development patterns in use. - Specific assumptions are described where any non-standard methods are in use or where there issues such as PPOs. - Detailed assumptions are provided relating to large loss estimates. - Numerical information (e.g. net:gross ratios) is provided relating to the reinsurance calculations together with reasoned argument supporting the choice of assumptions. - There is good, clear, description of assumptions relating to ULAE, bad debt, trust fund estimates and ultimate premiums and those assumptions are justified where justification would be appropriate. Key assumptions are described adequately and you have a good but not detailed understanding of them. In particular: - For claims projections there is good detail of assumptions (e.g. IELRs or ULRs) at a class by class level although there could perhaps be more detail. - There is some detail of assumptions relating to large loss estimates. - There is good description of assumptions relating to ULAE, bad debt, trust fund estimates and ultimate premiums. Assumptions are described in broad terms. The descriptions are such you only have a summary understanding of them and there is little information on assumptions relating to claims projections that is specific to classes of business. (5.8) How well does the report communicate the basis for the choice of the final results? Results - Method Selection Reasons for final selections are described well and you have a detailed understanding of the reasoning behind the selections. An excellent grading indicates both excellent description of how the results have been chosen and usually that both of the following are present: - Detail of approaches and why and how they have been used in particular circumstances - Detail of calculations associated with any bespoke analyses. Reasons for final result selections are described adequately and you have a good Reasons for final result understanding of the reasoning behind to the selections are described only selections. at a summary level. As a There may, however, be areas where additional result, you only have a highlevel understanding of the specific information about why some individual results or bespoke approaches have been reasoning behind to the selected would have been useful, or where selections. approaches to bespoke calculations could have been explained in more detail. Lloyd s 27
2014 SAO Review Process Range of scores over time Range of Results for all producers 100% Excellent 80% Good 60% Meets Requirements 40% Poor 20% Unacceptable 0% 2005 2006 2007 2008 2009 2010 2011 2012* 2013* *2012and 2013 Scores based on a 'Lite' review process Mean Maximum Score Minimum Score * 2012 and 2013 scores based on a Lite review process. Note: The two highest and two lowest scores from each year have been excluded from this analysis. Lloyd s 28
Executive Summary Nature & Accuracy of Data Methods Assumptions Emerging Experience Result Selection Results Presentation Uncertainty Business Knowledge Overall communication 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 2014 SAO Review Process All sections graded as meets requirements or above for all reports. Range of Results for all producers 0% Unacceptable Poor Meets Requirements Good Excellent Lloyd s 29
SAO Review Process going forwards We think: Process is worthwhile....but reduced benefit given as all good Propose: Lighter review only review circa 1/3 reports each year Feedback high level findings And individual results where applicable Lloyd s 30
Reminder: SAO Submission Dates YE 2014 submission dates Submission Deadline US Trust Fund SAOs 13 February 2015 Worldwide SAOs 19 February 2015 SAO Reports 31 March 2015 or earlier Please submit two copies of the SAO report one of which must be a hard copy, electronic copies encouraged Reports to Jerome Kirk, Market Reserving and Capital, Lloyd s, One Lime Street, EC3M 7HA Submit electronic copies via e-mail to SAOReports@lloyds.com Lloyd s 31
Large Loss Wordings Lloyd s 32
Large Loss Wordings As expected the number of large loss wordings 3 reduced significantly in 2013. 35 Number of large loss wordings 3 : 2013 vs 2012 30 25 20 15 10 5 0 Hurricane Sandy Other Subprime/ Credit Crunch Aggregated Wording WTC Thai Floods New Zealand Earthquakes 2011 Japanese Earthquake and Tsunami New Zealand Earthquake 2010 2012 2013 Source: Analysis of 2012 and 2013 SAOs Lloyd s 33
Producer 1 Producer 2 Producer 3 Producer 4 Producer 5 Producer 6 Producer 7 Total Producer 1 Producer 2 Producer 3 Producer 4 Producer 5 Producer 6 Total Large Loss Wordings - Quantification Information provided within the reports has improved however there is still some room for further enhancement 100% 80% 60% 2013 100% 40% 50% 50% 20% 0% 0% 0% 0% 28% 100% 80% 2012 100% 60% 40% 50% 55% 67% 50% 50% 20% 0% 0% 0% Lloyd s 34
Large Loss Wordings - Quantification Within the SAO report we continue to expect to see quantified justification for any large loss wording graded as 3 or 4 The appendix to this slide deck re-iterates what we mean The quantified deterioration of net reserves in a realistic adverse scenario is the most helpful approach. As we did this year we ll follow up when we need to The follow up information has been useful and timely but we d expect to see this information in the SAO report Lloyd s 35
Large Loss Wordings Clarification Following our request last year the clarity of the SAO large loss wordings has improved However we did still seek further clarification on aggregated wordings Aggregated wordings (across losses and/or year) are acceptable but please Clearly include the word aggregate Clearly identify which losses are being aggregated and for which years Please only refer to large loss wordings within the relevant SAO Please do not refer to a non-us WW wording in the US Trust Fund SAOs Lloyd s 36
Large Loss Wordings Proportionality We still get a handful of wording 3s where the quantified net unfavourable uncertainty looks very low or is zero For example : Aggregated wording 3 where on following up with the signing actuary the unfavourable net uncertainty was nil Wording 3 - an increase to the uncertainty of the syndicate s total reserves Please consider whether a wording 3 is really appropriate in these instances Lloyd s 37
Regulatory Developments Lloyd s 38
Valuation of Liabilities Rules - no significant changes this year Have been updated to include an additional requirement see paragraph 48 regarding the provision of information to Lloyd s Paragraph 48 Lloyd s expects all signing actuaries to provide Lloyd s with all relevant information documents and explanations, if requested. Where the work or results are being questioned Lloyd s may carry out additional investigations and we expect actuaries to cooperate fully with Lloyd s in these. Lloyd s 39
Valuation of Liabilities Additional Information Technical provisions for solvency best estimate basis reasonably foreseeable basis not the mean of all possible future outcomes not necessary to allow for emergence of new claim types or events not in data Bad Debt already covered but there is an FAQ now on this topic Lloyd s 40
The PRA have written to firms on reserving The PRA has written to all industry CEOs and including managing agents reserve setting process needs to identify key issues and sources of material uncertainty strong governance around the reserve setting process needs to be demonstrated reserve strength needs to be monitored and reported to the board assess whether Unexpired Risk Provision needs to be established have regard to underwriting, reserving and economic cycles ensure clear feedback loops between underwriting, claims and reserving have regard to data quality and consider how data limitations might impact firms reserves Lloyd s reserving framework and oversight (including minimum standards) covers this Lloyd s 41
Look out for APS X2 it will impact SAOs (and more ) Members must consider whether to apply Peer Review to Actuarial Work for which they are responsible. The extent of the peer review should have regard to a number of points including.. Significance of the work Reasonable expectation of the person for whom the work is produced for Difficulty/complexity of the work A clear understanding of roles is required, the timing should allow for the peer reviewer to have influence on the work Peer reviewers must be objective and have the appropriate experience and expertise Lloyd s 42
Future of SAOs - to be retained in PRA handbook PRA covered this in CP16/14 The PRA has considered whether responsibility should rest with the Society in the manner suggested by respondents and has decided to retain its current approach. This is to ensure that the PRA has a full range of tools to deploy in its supervision of managing agents. Therefore, the PRA will maintain a rule which requires the actuarial function of each managing agent carrying out general insurance business to review and provide an SAO to both the managing agent and the Society of Lloyd s regarding the technical provisions held in respect of each syndicate. Although still to be decided we expect: SAO on technical provisions for solvency (maybe excluding risk margin) Still one way test Can be provided internally or externally Lloyd s 43
Approved Persons - extended to actuarial function PRA released CP26/14 on November 24 th 2014 update of approved person regime Introduction of Chief Actuary Function (SIMF20) will apply to Lloyd s and Managing Agents Actuarial profession linking to upgrade of Practising Certificate Lloyd s 44
Practicing Certificates - consultation just out Linked to PRA CP Proposal is if member of profession and taking up SIMF20 then will need practising certificate covers all areas for actuarial function specifically an including Lloyd s version Will retain an SAO only version to allow flexibility in who provides SAO Life syndicate can use a life certificate as an alternative Lloyd s 45
Update on Actuarial Function Reports. Lloyd s has been reviewing Actuarial Function Reports throughout 2014 Submissions in May needed to cover TPs and Risk Management feedback provided to all agents in July subsequent re-submissions have been closing off issues and we are reviewing and re-issuing feedback as required We requested Underwriting and Reinsurance Opinions by the end of November reviews are ongoing feedback will be reissued once reviewed (by the end of the year at latest) Lloyd s 46
.some Early Feedback on Underwriting and Reinsurance Opinions Good progress has been made despite evolving requirements and expectations. Make sure: the opinions are clear and unambiguous. if improvements can be made (to processes or plans) then make the recommendations clear. if there are no concerns, state this too! all requirements are being met, if they are not material (e.g. Options and Guarantees) state this and give rationale. Reliance on work from others (e.g. internal model output) is acceptable and often very useful Lloyd s 47
Risk Margin PRA s Supervisory Statement 5/14 states that: firms should not approximate future Solvency Capital Requirements used to calculate the risk margin as proportional to the projected best estimate unless this has been shown not to lead to a material misstatement of technical provisions. Agents need to demonstrate this. For example: is reserve risk proportional to the reserves or does risk increase as the provisions get smaller? is operational risk really proportional to the technical provisions? Lloyd s 48
Risk Margin need to consider other approaches to estimating future SCRs To get a more risk based approach than the proportionate run-off of each element we considered: The square root of the claims run-off pattern A simplified Risk based approach assuming: Riskiness of net provisions and reinsurance recoveries increases by 10% a year 20% of operational risk relates to provisions and the risk on these reduces, all else is constant no unavoidable market risk or premium risk Lloyd s 49
Risk Margin - Results Square root of claims and risk run-off are very similar in the first 9 years when more than 90% of the provisions are paid. The square root approach seems to produce a reasonable match to the risk approach under our crude assumptions! You do need to justify this for you. Lloyd s 50
Contract Boundaries treatment of binders has been clarified Insurance or reinsurance contracts should be recognised as legal obligations at the earliest of the date the undertaking becomes party to the contract, or the date which cover begins Binding Authorities are not (re)insurance contracts Therefore a look-through, or a reasonable approximation is required to the underlying contracts This is clarified within the Level 3 Guidelines (para. 2.15 of Guidelines on Contract Boundaries) Approaches such as 3 month cancellation periods on Binding Authorities are no longer acceptable Lloyd s 51
Treatment of Reinsurance some clarification but some uncertainty remains Solvency II requires a cashflow approach in respect of current obligations Debate around the treatment of outwards premiums and claims current guidance is to use correspondence that is to only include elements of outwards reinsurance in TPs that relate to inwards bound contracts residual element of RI asset (or premiums payable) would sit outside TPs this does mean some cashflows would have been outside TPs EIOPA appear to disagree with concept of correspondence Lloyd s 52
Treatment of Reinsurance some clarification but some uncertainty remains Original Level 3 Guidelines Consultation (para 2.163) ruled out the splitting of the outwards premium between corresponding and noncorresponding inwards contracts. i.e. no splitting between TPs and Other Assets Uncertainty as to whether this implies that all cashflows from the outwards contract are to be allocated to TPs or no recognition is allowed for recoveries from inwards contracts not yet bound For 2014YE propose continue on current basis will seek further clarification and update our guidance in early 2015 (which will apply June 2015) Lloyd s 53
This means we will update the Solvency II Technical Provisions Guidance Will update in Q1 2015 To address: Risk Margin Methods Contract Boundaries Treatment of Reinsurance Binary Events/ENIDs Remember the audit of the QMC uses this as a reference document so it should be followed Lloyd s 54
Reminder: SAO Submission Dates YE 2014 submission dates Submission Deadline US Trust Fund SAOs 13 February 2015 Worldwide SAOs 19 February 2015 SAO Reports 31 March 2015 or earlier Please submit two copies of the SAO report one of which must be a hard copy, electronic copies encouraged Reports to Jerome Kirk, Market Reserving and Capital, Lloyd s, One Lime Street, London, EC3M 7HA Submit electronic copies via e-mail to SAOReports@lloyds.com Lloyd s 55
Questions Lloyd s 56
Lloyd s 57
Appendix Lloyd s 58
WHY ARE WE INTERESTED IN LARGE LOSS WORDINGS? The Lloyd s Actuary has to understand the uncertainty and to satisfy himself that the declared uncertainty is not material to Lloyd s globally To do this he relies on the syndicate statements of actuarial opinion. < Picture to go here > Large loss wordings are also used as part of our reserving oversight work 59 Lloyd s 59
HOW CAN WE UNDERSTAND THE LARGE LOSS WORDINGS? As requested at the 2012 Signing Actuaries Forum we expect to see quantified justification for any large loss wording in the SAO report Some of the reports included the required information, but not all..we had to follow up with the signing actuaries and in all cases we received useful responses We expect to have the quantification of uncertainty readily available and clearly sign-posted within the SAO report < Picture to go here > 60 Lloyd s 60
WHAT would we like to see In terms of quantification. The quantified deterioration of net reserves expected in a realistic adverse scenario (or similar) is the most helpful approach Realistic adverse scenario is something approaching a ~90 th percentile or similar Lloyd s would need to understand the rationale behind the selection of the realistic adverse scenario < Picture to go here > Potential deterioration Reserve 90th 61 Lloyd s 61
Large Loss Wordings: Potential EXAMPLEs The syndicate s net reserves to this event/loss are X.Xm, corresponding to an industry loss of $XXbn. An industry loss of $XXbn (which would be a realistic adverse scenario) would increase the syndicate s net reserves to approximately X.Xm. < Picture to go here > A realistic adverse scenario is one in which a deterioration to the XX reserves of XX% could be anticipated. This scenario would result in a deterioration to the syndicate s net reserves of X.Xm. 62 Lloyd s 62
WHAT we wouldn t like to see Quantification will be insufficient if Quantitative criteria shown but potential deteriorations not included Quantification is not readily available, e.g. based on exposure or benchmark which is not clearly stated An upper bound is provided which is much higher than the deterioration expected A range of potential deteriorations is provided and that range of potential deteriorations is large < Picture to go here > 63 Lloyd s 63
Regulatory Updates References Level 2/Delegated Acts: http://ec.europa.eu/finance/insurance/solvency/solvency2 /index_en.htm EIOPA Final Report on Consultation (set 1) on Solvency II Guidelines (Level 3): https://eiopa.europa.eu/consultations/consultationpapers/2014-closed-consultations/june-2014/publicconsultation-on-the-set-1-of-the-solvency-iiguidelines/index.html PRA Consultation Paper 26/14 on Senior insurance managers regime: http://www.bankofengland.co.uk/pra/pages/publications/c p/2014/cp2614.aspx < Picture to go here > Lloyd s 64
Lloyd s 65