Session 155 PD, Guaranteed Issue, Simplified Issue and Preneed Update. Moderator: Cynthia MacDonald, FSA, MAAA

Transcription

1 Session 155 PD, Guaranteed Issue, Simplified Issue and Preneed Update Moderator: Cynthia MacDonald, FSA, MAAA Presenters: David B. Atkinson, FSA Jeffrey E. Johnson, ASA, MAAA Lloyd M. Spencer Jr., FSA, CERA, MAAA SOA Antitrust Disclaimer SOA Presentation Disclaimer

2 Session 155 Guaranteed Issue, Simplified Issue and Preneed Update CINDY MACDONALD,FSA, MAAA - Moderator DAVID B. ATKINSON, FSA JEFFREY E. JOHNSON, ASA, MAAA October 26, 2016

3 Polling Question Warm-up I am attending this session because: A. I walked into the wrong session bye bye B. I heard the speakers were outstanding C. I heard free SOA pens were being given to all attendees D. I get really excited by mortality topics E. I m hungover from too much Las Vegas last night I figured this session would be ideal for a snooze F. I m not sure 46% 25% 12% 10% 7% 0% A. B. C. D. E. F. 2

4 Demographics Polling Question #1 Are you attending this session because you have an interest in: A. Guaranteed Issue? B. Preneed? C. Simplified Issue? D. Other? 71% 92% 40% Answer all that apply 8% A. B. C. D. 3

5 Demographics Polling Question #2 What kind of work do you do? A. Experience studies / research B. Product development C. Financial reporting / valuation D. Risk management E. Other 38% 61% 40% 17% 24% Answer all that apply A. B. C. D. E. 4

6 Data collected Observation years companies supplied data # of Companies Exposure Years (000,000) Death Claims (000) Exposure Units* (000,000) Death Units* (000,000) GI Preneed SI * Ultimate face amount divided by $1,000. 5

7 Update on Guaranteed Issue Mortality Table Development DAVID B. ATKINSON President, DB Atkinson Consulting October 26, 2016

8 Many thanks to Joe Hogan! Joe developed the Guaranteed Issue mortality table and created a model office for reserves Joe retired from New York Life last year but has continued to volunteer and work on GI; unfortunately, he is not attending this meeting Joe not only coached me on this presentation but Agreed to be my lifeline, so I can phone a friend if you ask a question I can t answer If I win the $1 million, I will share it with him! 7

9 Characteristics of the business studied Composite/unismoke only i.e., smoker/nonsmoker business was excluded Composite business had an average size of $6,500 Policy sizes of $25,000 and up were overwhelmingly smoker/nonsmoker, so most were indirectly excluded Below $25,000, 96% of business was direct marketed For $25,000 and up, 96% of business was sold through independent agents and brokers Results of these exclusions: Much more homogenous study data at a cost of excluding only 12% of exposure and 4% of death claims 4 of 15 data contributors were excluded from the study 8

10 GI Polling Question #1 Does your company issue GI business with characteristics similar to those for the GI mortality study (i.e., Direct Marketed, Modified DB, Unismoke, and Average Size under $10,000)? 45% 1. Yes, my company issues GI business with similar characteristics 2. No, my company issues GI business with different characteristics 3. My company does not issue GI business 30% 25%

11 GI Mortality Study Stats 4.8 million policy years exposed $31 billion of exposure Over 200,000 deaths Over $1.3 billion in death claims Average mortality rate of 4.3% Average issue age for exposure was 67 Average age at death was 70 10

12 GI Mortality Table Overview Five-year ANTI-select and ultimate table ( anti-s&u ) Data was sparse below age 50 and above age 85 Used ratios of GI to Preneed mortality to extend rates to youngest and oldest ages GI mortality rates roughly similar to Preneed in size and slope Preneed had more credible data at extreme ages GI mortality table validation: Ratios of actual to expected based on anti-s&u table were 100.2% by units, 98.9% by count. Ratios using ultimate-only GI table were about 4% higher Ultimate mortality was lower than anti-s&u mortality! 11

13 GI Anti-Select and Ultimate Table Female and Male Quinquennial Issue Ages 50, 55,, 80, 85 and Ultimate Ages Female Ultimate Male Ultimate Quinquennial Female Select Quinquennial Male Select 12

14 A/E Ratios to GI Ult-Only and GI anti-s&u Tables, by Contributing Company 300.0% A/E GI Ult 300.0% 250.0% A/E GI S&U 250.0% 200.0% 200.0% A/E Ratio 150.0% 100.0% 150.0% 100.0% 50.0% 50.0% 0.0% Contributing Company 0.0% 13

15 Coverage Ratios by Percentage of Companies and Exposure Covered After 6 of the 11 companies and 99.2% of exposure were covered, it required a large increase in loading to cover each additional company, while contributing a tiny amount of additional exposure. 117% of GI anti-s&u or 122% of GI Ult-Only covered 55% of companies and 99.2% of exposure 145% of GI anti-s&u or 160% of GI Ult-Only covered 64% of companies and 99.5% of exposure 157% of GI anti-s&u or 164% of GI Ult-Only covered 73% of companies and 99.9% of exposure 14

16 Model Office for GI Reserves The GI Basic Mortality Tables (Anti-Select & Ultimate, Ultimate-Only) were loaded using the same loading formula applied to 2014 VBT to produce 2017 CSO. Percentage loadings were: 22% at age 30 graded to 17% at age 80 graded to 15% at age 100 graded to 7.5% at age 110 Overall average loading was 17% 15

17 Model Office for GI Reserves (continued) Whole life product was modeled; representative of majority of business Joe built model office based on study data distributions by age group and sex. Model office projection was based on $1 million of initial annual premiums and one year of issue 16

18 Model Office Mean Reserves: 2001 CSO vs. GI Anti-S&U vs. GI Ult-Only 2,000 1,800 1,600 Mean Reserves ($000) 1,400 1,200 1, CSO Mortality 2015 GI Ultimate-Only Mortality 2015 GI Antiselect & Ultimate Mortality End of Year 17

19 Mean Reserve Comparison All figures were per $1 million of initial annual premium GI Ult-Only produced the largest mean reserves Even though its mortality rates were lower, GI Ult- Only produced larger reserves than GI Anti-S&U. Higher early year mortality rates for GI Anti-S&U reduced early year reserves 18

20 Mean Reserve Comparison GI Ult-Only mean reserves exceeded 2001 CSO mean reserves by: $250,000 (32%) after 3 years (this was the maximum) $127,000 (9%) after 6 years $57,000 (3%) after 9 years $20,000 (1%) after 12 years Starting in year 15, GI Ult-Only mean reserves exceeded 2001 CSO by no more than $4,000 19

21 Model Office Mid-Terminal Reserves: 2001 CSO vs. GI Anti-S&U vs. GI Ult-Only 2,000 1,800 Mid-Terminal Reserves ($000) 1,600 1,400 1,200 1, CSO Mortality 2015 GI Ultimate-Only Mortality 2015 GI Antiselect & Ultimate Mortality End of Year 20

22 Mid-Terminal Reserve Comparison Starting in year 7, 2001 CSO mid-terminal reserves were the largest of the three GI Ult-Only mid-terminal reserves exceeded 2001 CSO mid-terminal reserves by: $148,000 (25%) after 3 years (this was the maximum) $26,000 (2%) after 6 years Starting in year 7, 2001 CSO mid-terminal reserves exceeded GI Ult-Only mid-terminal reserves 21

23 Blending Mean and Mid-Terminal Reserves The distribution of GI business by premium mode was equivalent to 37% annual mode plus 63% monthly mode, measured by number of payments per year Mean reserves were used for annual mode Mid-terminal reserves were used for monthly mode Blended reserves were calculated as 37% mean reserves + 63% mid-terminal reserves 22

24 2,000,000 Blend of 37% Mean and 63% Mid-Terminal Reserves Blended Reserves ($000) 1,800,000 1,600,000 1,400,000 1,200,000 1,000, , , , , CSO Mortality 2015 GI Ultimate-Only Mortality 2015 GI Antiselect & Ultimate Mortality End of Year 23

25 Blended Reserve Comparison For the first 8 years, GI Ult-Only Blended Reserves were the largest of the three GI Ult-Only blended reserves exceeded 2001 CSO blended reserves by: $186,000 (28%) after 3 years (this was the maximum) $63,000 (5%) after 6 years Starting in year 9, 2001 CSO blended reserves exceeded GI Ult-Only blended reserves 24

26 Result of Presentation to LATF, August, 2016 GI Ult-Only tentatively chosen as basis for GI valuation and NFF. Male and Female Composite (unismoke) rates Much closer fit for GI business than any existing table Need to develop clear definition of what is and what is not guaranteed issue Need writers of GI business to review and test GI Ultimate-Only table 25

27 Next Steps Define GI Gather industry feedback Document new GI tables Likely effective date? 26

28 GI Polling Question #2 Which mortality table would you prefer future GI reserves to be based on: 50% CSO (default) CSO GI Ultimate- Only Table GI Anti-Select & Ultimate Table 24% 9% 17%

29 GI Polling Question #3 For NFF, would you prefer to use: CSO (which will apply to your non-gi business) 2. Same mortality table used for GI reserves 35% 65%

30 Your Thoughts?

31 Preneed Mortality Table Update Jeffrey E. Johnson, ASA, MAAA October 26, 2016

32 Agenda History & Background Data Findings Preneed experience table Comparisons to CSO tables Mortality Mean reserves LATF discussion and outcome Next steps 31

33 History 2008 preneed table 10 participating companies experience 1,708,924 lives exposed 143,422 deaths (8.4% mortality rate) Objective: Table for nonforfeiture, stat & tax valuation, IRC 7702 Result: Basic experience table; 1980 CSO adopted as official table for preneed in most states 32

34 Background Combined study commissioned by NAIC s Life Actuarial Task Force (LATF): Guaranteed Issue Simplified Issue Preneed Data gathered & processed by MIB (2011) Objective: Table for nonforfeiture, stat & tax valuation, IRC 7702 Byproduct: Industry experience 33

35 Data Data from calendar years contributing companies 7.9 million policy years exposed 635,000 death claims $35 billion exposed $3.0 billion of claims Single premium policies accounted for approximately 60% of exposure and 75% of deaths 96% issued on unisex basis 34

36 Data - Segmentation Sex Male, Female, Unisex Age Basis Near, Last Birthday Underwriting Guaranteed, Simplified DB Pattern Modified, Level, Increasing Premium Pattern Single, Limited (Multi) Distribution Independent Agents, Other Applicant Type Individual, Association Grp Face Amount Under $10k, $10k & greater 35

37 Polling Question #1 Preneed applications for the company(s) I work for are underwritten: A. As guaranteed issue B. Using only responses to a limited set of broad health questions C. Sometimes using supplemental data (e.g., MIB or prescription drug checking services) D. Sometimes using more traditional techniques (e.g., blood, urine, paramed exam, APS, etc.) 72% 22% 6% 0% A. B. C. D. 36

38 Polling Question #2 The preneed company(s) I work for sell deferred annuities for the purpose of: A. We don t sell annuities B. Handling imminent deaths C. Providing flexible payment arrangements D. Other reasons 30% 25% 25% 20% A. B. C. D. 37

39 Polling Question #3 The preneed company(s) I work for offer payment plans that extend to only: A. 10 years B. 15 years C. 20 years D. Life Pay 71% 24% 0% 6% A. B. C. D. 38

40 Findings Mortality varied significantly by segment Mortality was much less volatile by company when all segments were combined Consistent with similar overall risk pools for each company Evidence that companies coded SI vs. GI differently could account for variations by product/underwriting segment Anti-select mortality No evidence of mortality improvement 39

41 Development of Preneed Experience Separate rates for unisex and male/female, graduated in 5-year issue age groups, 10-year select and ultimate Select mortality was generally anti-select, especially policy year 1 For issue ages 0-64, mortality became select around policy year 5 Rates for issue ages under 40 were smoothed: Overall data for ages 0-39 were significant but very lumpy by duration Rates for attained ages 97+ set consistently higher than 2015 VBT, grading to a rate of for attained ages 110+ Developed rates for individual issue ages

42 Preneed Experience Table Sex Unisex, Female (63%), Male (37%) 3 segments (pay plans / risk classes) Single Pay (guaranteed issue, level benefit) Multi-Pay (guaranteed issue, graded benefit) Multi-Pay (simplified issue, level benefit) 41

43 Preneed Experience Table (Unisex) Duration 1 masks later duration mortality (anti-selection) Mortality Rate Issue Age 42

44 Mortality Comparison Issue Age 65 CSO tables are ultimate rates only, unisex 60% female and 40% male Floored preneed mortality rates are lesser of Basic Preneed and 2017 CSO rates 2015 Preneed table is not loaded 43

45 Mortality Comparison Issue Age 85 CSO tables are ultimate rates only, unisex 60% female and 40% male Floored preneed mortality rates are lesser of Basic Preneed and 2017 CSO rates 2015 Preneed table is not loaded 44

46 Year 1 Mean Reserve Impact Estimated Year 1 mean reserve increase per $1000 face amount Calculated using 2015 Preneed Basic Mortality, CRVM method, 3.50% interest Single Pay Issue Age Year 1 Mean Reserve Change Relative to 1980 CSO-E 2015 Preneed Basic 2017 CSO Composite Ultimate 65 -$6 -$ $6 -$63 85 $0 -$39 Ten Pay Full DB Issue Age Year 1 Mean Reserve Change Relative to 1980 CSO-E 2015 Preneed Basic 2017 CSO Composite Ultimate 65 +$10 -$6 75 +$15 -$8 85 +$29 -$17 10-pay projected mean reserves increase with the 2015 Preneed Basic Mortality. This is due to the reverse select and ultimate structure resulting in an expense allowance of $0. 45

47 Mean Reserve Comparison Issue Age 65, Unisex, Single Pay CSO tables are ultimate rates only, unisex 60% female and 40% male Floored preneed mortality rates are lesser of Basic Preneed and 2017 CSO rates 2015 Preneed table is not loaded 46

48 Mean Reserve Comparison Issue Age 85, Unisex, Single Pay CSO tables are ultimate rates only, unisex 60% female and 40% male Floored preneed mortality rates are lesser of Basic Preneed and 2017 CSO rates 2015 Preneed table is not loaded 47

49 Mean Reserve Comparison Issue Age 65, Unisex, 10 Pay Full DB CSO tables are ultimate rates only, unisex 60% female and 40% male Floored preneed mortality rates are lesser of Basic Preneed and 2017 CSO rates 2015 Preneed table is not loaded 48

50 Mean Reserve Comparison Issue Age 85, Unisex, 10 Pay Full DB CSO tables are ultimate rates only, unisex 60% female and 40% male Floored preneed mortality rates are lesser of Basic Preneed and 2017 CSO rates 2015 Preneed table is not loaded 49

51 Valuation Loading Analysis Contributing companies represent >65% of 2015 preneed sales Overall Preneed mortality was 99.8% of the 2015 Preneed Table Preneed Loadings: -0.2% loading needed to cover 70% of contributing cos. +2.3% loading needed to cover 80% of contributing cos. +3.9% loading needed to cover 90% of contributing cos. 50

52 Guidance Requested from LATF Which table to use for valuation and nonforfeiture? 2015 Preneed Basic table select & ultimate plus loads what loads? 2017 CSO Ultimate Continued use of 1980 CSO preferred by carriers 51

53 LATF Decision Continue using 1980 CSO Continue PBR exclusion Endorse 2015 Preneed Basic experience table 52

54 Not All States Recognize 1980 CSO 10 outstanding states DE, DC, GA, HI, MA, MS, MT, PA, VT, WA Lack of uniformity creates product filing and administration pain If 1980 CSO is not recognized as the minimum basis for preneed in these states, 2017 CSO will become the minimum basis in

55 Next Steps Document development of 2015 Preneed Basic Tables Work with states to achieve 100% recognition of 1980 CSO minimum standard for preneed 54

56 Simplified Issue & Accelerated Underwriting Session 155 Panel Discussion - GI / SI / Preneed Update Lloyd M. Spencer, Jr., FSA, CERA, MAAA SOA Annual Meeting & Exhibit Las Vegas, NV - October 26, 2016

57 Agenda Simplified Issue Mortality Table Development Update A single table or approach is difficult for Simplified Issue business and a single SI mortality table no longer makes sense A New Approach Accelerated Underwriting Identifying Accelerated Underwriting Next Steps

58 Simplified Issue Mortality Table Development Update NAIC s Life Actuarial Task Force (LATF) Charge Simplified Issue study conducted by the Joint AAA Life Experience Committee and the SOA Preferred Mortality Oversight Group Data collected from issue years 30 contributing companies 26 companies contributed non-smoker / smoker data 18 companies contributed uni-smoke data Over $490 billion exposed on just under 19 million policy years $2.3 billion in death benefits from just over 270,000 claims Data limitations Majority of data collected is from policies with smaller face amounts and on a unismoke basis Non-smoker / smoker claims data much less robust and more disparate across contributing companies Limited older issue and attained age data Data gathered insufficient to allow study by distribution method or target market of underlying business

59 SI Mortality Table Development Update (cont'd) A/E Ratio of Experience (expected = 2008 VBT Limited Underwriting Table) Data by Contribution Segment 2008 VBT Limited Underwriting Table UW Class # of Claims S&U Ultimate Non-smoker 35, % 86% Smoker 16, % 92% Uni-smoke / Composite 220, % 115% Total 272, % 101% Source: Joint AAA/SOA SI Study Committee Data Grouping Issue Amounts # of Deaths Average Issue Amount By Band All 1. Uni-smoke / Composite < $50, ,704 $ 5, Uni-smoke / Composite > $50,000 1,215 $52, Non-smoker / Smoker < $50,000 46,904 $12, Non-smoker / Smoker > $50,000 4,854 $81,247 Source: Joint AAA/SOA SI Study Committee $ 6,053 $18,624

60 SI Mortality Table Development Update (cont'd) New Simplified Issue experience mortality table developed for uni-smoke only business with issue amounts under $50,000 Characteristics of the table Separate rates by gender 10-year select and ultimate Younger issue ages had scant data Core ages follow a pattern fairly consistent with the SOA Basic Mortality Table Despite the good overall table fit for this segment, loads to enable coverage of 60% to 80% were significant, driven by the variability of experience across Contributing company Issue age Gender

61 A single table or approach is difficult for SI business... Fully Underwritten Range of Mortality Experience Simplified Issue Mortality outcome for any underwriting regime is a factor of many selection levers Two similar underwriting programs will not necessarily result in a similar experience mortality outcome Key determinants to experience mortality outcome include: Structure of application Target market of product Distribution channel Individual selection criteria Company s claims adjudication policies and processes Company s management of the product

62 ...and a single SI mortality table no longer makes sense Pricing and underwriting have advanced Prices have improved with lowering mortality through the use of new underwriting tools, such as: Prescription database checks Motor vehicle reports Medical Information Bureau (MIB) checks Credit / financial information checks Underwriting algorithms Much more business issued on a non-smoker / smoker basis Significant increase in the size of policy face amounts issued Clear definition of what business is SI and what business is not SI Definition of SI business complicated with emergence of new, instantly-available data sources, as well as process automation

63 A New Approach - Accelerated Underwriting Accelerated Underwriting Dynamic process that customizes the volume and type of non-medical & medical information gathered for each individual applicant May result in different data being evaluated for two policies issued from the same product, at the same face amount, and with the same gender, issue age and tobacco status Impact on retail premium closer to traditional fully-underwritten business than SI business Approaches vary, and rely upon one or more of the following: Use of both traditional and non-traditional data sources; Use predictive models to quickly interpret available information; and Separate applications into those that may be rated using non-traditional methods v. traditional U/W Goal of Accelerated Underwriting is to right size the underwriting approach without increasing the likelihood that the collection of additional information would change the underwriting decision

64 Identifying Accelerated Underwriting May look like an expanded SI process, with mortality expectations that align more closely with fully-underwritten business Often (though not always) uses predictive modeling and other complex algorithms May also include: The use of traditional underwriting sources, collected through different means (MIB, MVR, criminal history, prescription database checks electronic lab data, electronic health records An expanded or contracted application combined with a tele-underwriting / interview process Non-traditional data, such as clinical laboratory data, credit profiles, facial analytics, etc. Would often exclude the use of fluids (blood & urine), for those cases that qualify Approaches can and will be different, and will produce different expected mortality and a shifting mix of business between preferred and standard underwriting classifications

65 Next Steps Joint AAA/SOA Committee recommendation to LATF (August 24 Meeting) Finalize the development and expose the new uni-smoke, smaller issue amount SI experience mortality table Request industry testing of new table against more recent business Develop a definition of SI business that would use this table Request additional guidance from LATF regarding a new approach to developing valuation mortality for SI business and Accelerated Underwriting business Develop a definition of such business and address treatment under Principles Based Reserving, including consideration of companies determination of the following parameters under VM-20 Applicable industry table Application for mortality segment Margin level

66 Disclaimer The information provided in this presentation does in no way whatsoever constitute legal, accounting, tax or other professional advice. While Hannover Rück SE has endeavoured to include in this presentation information it believes to be reliable, complete and up-to-date, the company does not make any representation or warranty, express or implied, as to the accuracy, completeness or updated status of such information. Therefore, in no case whatsoever will Hannover Rück SE and its affiliated companies or directors, officers or employees be liable to anyone for any decision made or action taken in conjunction with the information in this presentation or for any related damages. Hannover Rück SE. All rights reserved. Hannover Re is the registered service mark of Hannover Rück SE SOA Annual Meeting & Exhibit - Session October 26, 2016