Policyholder Behavior Experience Data and Modeling Timothy Paris - - PDF document

Equity-Based Insurance Guarantees Conference

• Nov. 11-12, 2019

Chicago, IL

Policyholder Behavior Experience Data and Modeling Timothy Paris

SOA Antitrust Compliance Guidelines SOA Presentation Disclaimer

Sponsored by

SO SOA E Equit ity-Base sed I d Insur uranc nce G e Guarantees es C Conf nfer erenc nce Policyho holder der Beha havior E Expe perienc ence D Data a and M d Modeling ng

Sessio ion 2B 2B TIMOTHY PARIS, FSA, MAAA RUARK CONSULTING, LLC

November 11, 2019 1:30 - 3:00pm Central

SOA A Antitrust C t Compliance G e Guidel elines es

Active participation in the Society of Actuaries is an important aspect of membership. While the positive contributions of professional societies and associations are well-recognized and encouraged, association activities are vulnerable to close antitrust scrutiny. By their very nature, associations bring together industry competitors and other market participants. The United States antitrust laws aim to protect consumers by preserving the free economy and prohibiting anti-competitive business practices; they promote competition. There are both state and federal antitrust laws, although state antitrust laws closely follow federal law. The Sherman Act, is the primary U.S. antitrust law pertaining to association

• activities. The Sherman Act prohibits every contract, combination or conspiracy that places an unreasonable restraint on trade. There are, however, some activities that are illegal

under all circumstances, such as price fixing, market allocation and collusive bidding. There is no safe harbor under the antitrust law for professional association activities. Therefore, association meeting participants should refrain from discussing any activity that could potentially be construed as having an anti-competitive effect. Discussions relating to product or service pricing, market allocations, membership restrictions, product standardization or other conditions on trade could arguably be perceived as a restraint on trade and may expose the SOA and its members to antitrust enforcement procedures. While participating in all SOA in person meetings, webinars, teleconferences or side discussions, you should avoid discussing competitively sensitive information with competitors and follow these guidelines:

• Do n
• not
• t discuss prices for services or products or anything else that might affect prices
• Do n
• not
• t discuss what you or other entities plan to do in a particular geographic or product markets or with particular customers.
• Do n
• not
• t speak on behalf of the SOA or any of its committees unless specifically authorized to do so.
• Do

Do leave a meeting where any anticompetitive pricing or market allocation discussion occurs.

• Do

Do alert SOA staff and/or legal counsel to any concerning discussions

• Do

Do consult with legal counsel before raising any matter or making a statement that may involve competitively sensitive information. Adherence to these guidelines involves not only avoidance of antitrust violations, but avoidance of behavior which might be so construed. These guidelines only provide an overview

• f prohibited activities. SOA legal counsel reviews meeting agenda and materials as deemed appropriate and any discussion that departs from the formal agenda should be

scrutinized carefully. Antitrust compliance is everyone’s responsibility; however, please seek legal counsel if you have any questions or concerns.

2

Presentation Disclaimer

Presentations are intended for educational purposes only and do not replace independent professional judgment. Statements of fact and opinions expressed are those of the participants individually and, unless expressly stated to the contrary, are not the opinion or position of the Society of Actuaries, its cosponsors or its committees. The Society of Actuaries does not endorse

• r approve, and assumes no responsibility for, the content, accuracy or completeness of the

information presented. Attendees should note that the sessions are audio-recorded and may be published in various media, including print, audio and video formats without further notice.

3

Background

4

Wh Why i y is this imp mpor

• rtant?
• Critical risk element for long-term EBIG-type products, and significant interactions

with capital markets risks

• Complex and dynamic data
• Data sparsity, especially at the company level – a credibility problem
• Data is emerging in key areas
• Asset-side folks – demand that your liability-side colleagues demonstrate the

robustness of models/assumptions that they provide you

5

Industr try s studies es

Fixed-indexed annuity policyholder behavior https://ruark.co/ruark-releases-2019-fixed-indexed-annuity-study/ https://ruark.co/ruark-consulting-releases-2018-fixed-indexed-annuity-mortality-study/ Variable annuity policyholder behavior https://ruark.co/ruark-releases-2019-variable-annuity-study-results/ https://ruark.co/ruark-consulting-releases-variable-annuity-mortality-study-results/

6

VM VM-21 PBR f for V Variable e Annuities es

Public redline exposure draft as of April 30, 2019 https://naic-cms.org/exposure-drafts Section 10: Contract Holder Behavior Assumptions Should examine many factors including cohorts, product features, distribution channels, option values, rationality, static vs dynamic Required sensitivity testing, with margins inversely related to data credibility Unless there is clear evidence to the contrary, should be no less conservative than past experience and efficiency should increase over time Where direct data is lacking, should look to similar data from other sources/companies 1 2 3 4

7

You and your data

8

0% 35% 7 or more 6 5 4 3 2 1

• 1
• 2
• 3 or

more

Years Remaining in Surrender Charge Period 2008 2008 2016 2016 2018 2018

Your company-level data might indicate some key patterns in surrender behavior

9

GLWB

Surrender Rate

9

0% 30% 7 or more 6 5 4 3 2 1

• 1
• 2
• 3 or

more

Years Remaining in Surrender Charge Period

Surrender rates are lower with living benefits…

10

Surrender Rate None ne GLW LWB Hybr brid id G GMI MIB

10

0% 25% 7 or more 6 5 4 3 2 1

• 1
• 2
• 3 or

more

Surrender Rate Years Remaining in Surrender Charge Period

GLWB - Withdrawal Behavior

11

…and even lower with income utilization

No prior WD WDs Ex Excess ss W WDs Le Less than or

• r fu

full W WDs Ds

11

…and when guarantees are more valuable

12

GLWB (nominal moneyness basis)

0% 25% 7 or more 6 5 4 3 2 1

• 1
• 2
• 3 or

more

Surrender Rate Years Remaining in Surrender Charge Period

ITM 50+% ITM 25 - 50% ITM 5 - 25% ATM OTM

12

Dynamic sensitivity has also changed over the years

13

0% 35% 3Q 09 3Q 10 3Q 11 3Q 12 3Q 13 3Q 14 3Q 15 3Q 16 3Q 17 3Q 18

GLWB Shock Lapse

ATM <25% ITM 25%-50% ITM 50%-100% ITM

Surrender Rate

13

How you measure value matters, but company-level credibility is very limited

14

0% 25% OTM 50+% OTM 25 - 50% OTM 5 - 25% ATM ITM 5 - 25% ITM 25 - 50% ITM 50 - 100% ITM 100%+

Surrender Rate

GLWB

Shock k - actuar arial Ul Ulti timate te - actuar arial Shock - nominal Ultimate - nominal

14

Largest and smallest contracts behave differently

15

0% 20% 7 or more 6 5 4 3 2 1

• 1
• 2
• 3 or

more

Surrender Rate Years Remaining in Surrender Charge Period

under 50,000 50,000-100,000 100,000-250,000 250,000-500,000 500,000-1,000,000 >=1,000,000

15

Withdrawals vary by age and tax status

16

Qualified Non-Qualified

0% 100% <50 50-59 60-64 65-69 70-79 80+

Frequency Attained Age

GLWB

16

Withdrawal behavior is becoming more efficient

17

0% 50% Q1 2008 Q1 2009 Q1 2010 Q1 2011 Q1 2012 Q1 2013 Q1 2014 Q1 2015 Q1 2016 Q1 2017 Q1 2018

Frequency

GLWB

LT Full WDs Full WDs Excess WDs

17

0% 10% <50 50-59 60-64 65-69 70-79 80-LEA Last Eligible

Annuitization Rate

Hybrid GMIB annuitization rates are low, but company-level credibility is very limited

18

18

2012 IAM does not fit VA mortality experience very well

19

0% 150% 0-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90+

% of Table

Actual vs. 2012 IAM - Projection G2

Male Count Male Amount Female Count Female Amount Base

19

Evidence of anti-selection for death benefit guarantees

20

0% 200% 1 2 3 4 5 6 7 8 9 10 11+

% of Table Dur uratio ion

Actual vs. 2012 IAM-G2

LB No LB

20

Results vary o

• ver

er t time a e and bet etween en c companies es

• Each company’s size affects quality of analytical insights and volatility of their
• wn results – a credibility problem
• Composition differences
• Idiosyncratic differences – product features, distribution, closed blocks, etc
• Using only your data, it is very difficult to identify the signal from the noise

21

Building models with your data

22

Model eling g and a assumptions

• Measuring goodness-of-fit for candidate models
• Testing predictive power on out-of-sample data
• Art + science: choosing, communicating, and ongoing recalibration

23

Goodness

• f Fit

Predictive Power

24

1,000 2,000 3,000 4,000 1 2 3 4 5 6 7 8 9 10

Bayesian Information Criterion (BIC) Number of Factors

25

0% 2% 4% 6% 8% 10% 1 2 3 4 5

Coefficient Standard Error Factor

26

27

0.00% 1.00% 2.00% 3.00% 1 2 3 4 5 6 7 8 9 10

Avg Abs A/E Error Number of Factors

28

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

True Positive Rate False Positive Rate

29

20,000 40,000 60,000 80,000 100,000 120,000 140,000 160,000 180,000 95% 96% 97% 98% 99% 100% 101% 102% 103% 104% 105% 1 2 3 4 5 6 7 8

Actuals A/E Factor Xi

30

• 40%
• 30%
• 20%
• 10%

0% 10% 20% 30% 40% 10 20 30 40 50 60 70 80 90 100

Actual to Model Average Expected Deciles

31

Cost-benefit of industry data

32

Ex Example: variable annuity ty i industry d data

• 24 companies
• Seriatim monthly data for policyholder behavior and mortality
• January 2008 through December 2018
• $795 billion ending account value

33

How you measure value matters, and credibility is vastly improved with industry data

34

0% 25% OTM 50+% OTM 25 - 50% OTM 5 - 25% ATM ITM 5 - 25% ITM 25 - 50% ITM 50 - 100% ITM 100%+

Surrender Rate

GLWB

Shock k - actuar arial Ul Ulti timate te - actuar arial Shock - nominal Ultimate - nominal

34

Industry data shows that GLWB income commencement is highest at issue and after bonuses expire…

35

0% 30% 1 2 3 4 5 6 7 8 9 10 11 12 13

Duration Frequency

35

0% 40% ITM 100+% ITM 50 - 100% ITM 25 - 50% ITM 5 - 25% ATM OTM 5 - 25% OTM 25+%

Frequency

…and that ultimate income commencement is dynamic

36

Dur 11+ r 11+ Dur 3 r 3-10 10

36

0% 10% <50 50-59 60-64 65-69 70-79 80-LEA Last Eligible

Annuitization Rate

Industry data shows that hybrid GMIB annuitization rates are backloaded…

37

37

…and depend on economic value of other benefits, such as continued income utilization

38

0.0% 2.0% <95% 95-100% 100-105% 105-110% 110-115% >115%

Annuitization Rate Ratio of Income PV to Annuitization PV

38

Industry data also makes a better tabular mortality basis…

39

39

0% 20% 40% 60% 80% 100% 120% 140% 0-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85-89 90+

% of Table

Actual vs. Ruark VAM 2015

Male Count Male Amount Female Count Female Amount Base

…and shows how income utilization affects mortality

40

0% 150% First Year No Prior Withdrawals Prior LT and/or Full WDs only Any Prior Excess WDs

% of Table

Actual vs. RVAM 2015

Qua ualif lifie ied No Non-qua ualif lified ied Total

40

Model eling g and a assumptions

• Measuring goodness-of-fit for candidate models
• Testing predictive power on out-of-sample data
• Using

ng r relevant indus dustry d data t to i improve ve c candi dida date m models

• Art + science: choosing, communicating, and ongoing recalibration

41

1,000 2,000 3,000 4,000 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Bayesian Information Criterion (BIC) Number of Factors

Company-only Industry

42

0% 2% 4% 6% 8% 10% 1 2 3 4 5 6 7 8 9 10

Coefficient Standard Error Factor

Company-only Industry

43

44

0.00% 1.00% 2.00% 3.00% 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Avg Abs A/E Error Number of Factors

Industry Company-only

Customize e your m model el i in a cred edibility-based ed f framework

• Subject matter expertise
• Actuarial judgment
• Quantify the benefits of using relevant industry data
• Ongoing recalibration, so focus on the framework and its sense of range

45

46

0% 2% 4% 6% 8% 10% 1 2 3 4 5 6 7 8 9 10

Coefficient Standard Error Factor

Industry Customized ed b blend Company-only

47

0.00% 1.00% 2.00% 3.00% 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Avg Abs A/E Error Number of Factors

Company-only Industry Customized ed b blend

How m w much i is 1% A/E i E improvem emen ent w worth th to to y you?

Suppose 5.00% average annual surrender rates for your block 1% A/E improvement would be 0.05% annually and about 0.60% in present value terms With 15% annualized market vol, hedge breakage (~2 s.d.) would be 0.18% of notionals So what are your hedge notionals?

48

Hedg edge n e notionals Annu nnualized ed hedge ge break akage ( (~ 2 2 s.d.) $100 million $180,000 $1 billion $1,800,000 $10 billion $18,000,000

0.60% * 15% * 2

Cost st-ben enefi fit o

• f i

industry d data

• Need to customize your model in a credibility-based framework
• Quantify the improvement in goodness-of-fit and predictive power metrics
• Quantify these improvements in financial terms – pricing margins, reserves, hedge

breakage

• Quantify the cost to access and use relevant industry data
• Altogether, does this improve your financial risk profile?
• Cont

ntrast t this a approach w h with h unl unlocking a ad na d naus useam

49

More data and/or relevant industry data Art + science, subject matter expertise and actuarial judgment More statistically justifiable model factors and dramatically improved fit and predictive power

50

Discussion

51