copy 2020 American Academy of Actuaries All rights reservedcopy 2020 Society of Actuaries All rights reservedMay not be reproduced without express permission

LIFE MORTALITY IMPROVEMENT SUBGROUP (LMISG) 2020 LIFE MI SCALE RECOMMENDATION

Life Actuarial (A) Task Force ndash NAIC Summer National Meeting 2020

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

2 copy 2020 Society of Actuaries All rights reserved

Agenda for Discussion

1 Objective for the annual mortality improvement (MI) scale updates

2 Current annual update methodology (2013ndash2019)3 Considerations for 20204 LMISG 2020 recommendation5 Future issues

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

3

Objective of Annual MI Scale UpdatesAddresses VM20 Incorporation of MI Section 9C3g

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

4

Objective of Annual MI Scale UpdatesLevel of Event Covered ndash Reserve vs Capital VM Introduction

Source VM20 ndash guidance provided on risks to be covered by reserves

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

5

Objective for Annual MI Scale UpdatesLMISG Thinking

Our annual update exercise seeks to apply judgment to historical mortality improvement (or deterioration) data to arrive at a set of mortality rates that can be used to calculate reserves for future events

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

6 copy 2020 Society of Actuaries All rights reserved

Most recent relevant historical MI data (10-year moving average)Historical Data

bull Age- and gender-basedbull Use of a long-term consistent source of population data Social Security Administration (SSA)

Most recent forecast of future improvements over future period (20 years)Forecasted Expectations

bull Age- and gender-based databull Consistent with historical data and projections (Alt II) available from SSA Trusteesrsquo Annual Report

Weighted average of historical data and forecasted expectationsUnsmoothed MI Scale

bull Average of historical data and forecasted expectations

Unsmoothed MI scale with smoothing process appliedSmoothed MI Scale

Current annual update methodology (2013-2019)

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

7

Preliminary 2020 ndash Current MethodologyBefore manual smoothing

-150

-100

-050

000

050

100

150

200

250

300

0 8 16 24 32 40 48 56 64 72 80 88 96 104

112

Male

Male 2020 Male 2019

-150

-100

-050

000

050

100

150

200

250

300

0 7 14 21 28 35 42 49 56 63 70 77 84 91 98 105

112

119

Female

Female 2020 Female 2019

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

8

Considerations for 2020 amp beyondbull Data - we donrsquot have sufficient data to fully understand the impact of the COVID-19-related mortality shock on

the insured population (anecdotal reports from companies indicate they are seeing a smaller shock)

bull MI scale updates - reflecting a shock in 2020 does not seem in line with the goals for the MI scale updates bull Are shock events more appropriately reflected in capital planning rather than reservesbull An effective vaccine may make COVID-19rsquos impact on MI much shorter than the long-term impact arising from the opioid

epidemicbull However COVID-19 may have potential longer-term impacts that may arise from survivor impaired health health

impacts from delays in health care andor testing for early detection of dread disease etcbull Conversely some experts and models indicate the 20202021 COVID-19 shock is mainly a moving forward of deaths that

would have occurred due to other causes andor comorbidities Might that improve future mortality improvement

bull Precedent for other excess mortality events bull First group to consider the impact of a short-term shock event ndash setting a precedent for other future MI scale workbull The current methodology uses a moving average to ldquosmooth outrdquo the impact of any one year or event bull 20082009 influenza season and the effect of the opioid epidemic ndash the methodology was not adjusted for those events

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

9

LMISG 2020 Recommendation

Apply the historical methodology for 2020 consistent with the past scale updates (2013ndash2019)

Implicationsbull There will be no specific impact included for the 2020 scale for the pandemic shock effectbull The 10-year historical average in the 2022 scale update will include a ldquosmoothedrdquo impact

of the shock as part of the usual methodology

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

10

Future Issuesbull Insured vs general population impacts

bull Some evidence that impact on insured population will be lessbull SOA ldquosocioeconomic decilerdquo study will provide some guidance herebull Consideration of consistent framework and changes to the current methodology (ex

averaging periods)

bull Will COVID-19 have a long-term impact on mortality improvement rates and what will the impact be

bull Lower due to survivor impaired health as well as the indirect effect arising from the virus delaying the early condition diagnosis of dread diseases and preventive treatments

bull Higher due to greater application of good hygienic habits (eg social distancing and washing hands) andor higher utilization of other vaccines (such as the annual flu shot)

bull Need to understand the impact in terms of potential effects on future slope and size of MIbull Impact in light of a COVID-19 vaccine availability and effectiveness

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

copy 2020 Society of Actuaries All rights reserved

Questions

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

copy 2020 Society of Actuaries All rights reserved

Appendix

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

13 copy 2020 Society of Actuaries All rights reserved

General PopulationPattern of excess deaths vs general mortality

AE Ratios ndash COVID-19 vs 2015 VBT Actual deaths = COVID-19

Expected deaths = 2015 Unismoke ANB 2015 VBT rates weighted by 72019 estimated US population

Slope of VBT matches COVID-19 female rates better than male rates

Small ratios gtgtgt only 1 cause of death (COD) in numerator all CODs in denominator

0904

0209

2731

4650

464647

46

1103

0506

2541

7580

6857

4657

lt1

5ndash14

25ndash34

45ndash54

65ndash74

85+

1ndash4

15ndash24

35ndash44

55ndash64

75ndash84

All Ages

Age

Grou

p

US Male

US Female

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

14 copy 2020 Society of Actuaries All rights reserved

2013-2019 MI Scale Update Reports2019 Scale httpswwwsoaorgresourcesexperience-studies2019mortality-improvement

2018 Scale httpswwwsoaorgresourcesexperience-studies20182018-mortality-improvement

2017 Scale httpswwwsoaorgresourcesexperience-studies2017mortality-improvement-2017

2016 Scale httpswwwsoaorgresourcesexperience-studies2016research-mortality-improvement-2016

2015 Scale httpswwwsoaorgresourcesexperience-studies2015research-2015-mortality-improvement-rate-year-end

2014 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-improve-rates-ag-38-year-end-2014

2013 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-imp-rates

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

2 copy 2020 Society of Actuaries All rights reserved

Agenda for Discussion

1 Objective for the annual mortality improvement (MI) scale updates

2 Current annual update methodology (2013ndash2019)3 Considerations for 20204 LMISG 2020 recommendation5 Future issues

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

3

Objective of Annual MI Scale UpdatesAddresses VM20 Incorporation of MI Section 9C3g

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

4

Objective of Annual MI Scale UpdatesLevel of Event Covered ndash Reserve vs Capital VM Introduction

Source VM20 ndash guidance provided on risks to be covered by reserves

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

5

Objective for Annual MI Scale UpdatesLMISG Thinking

Our annual update exercise seeks to apply judgment to historical mortality improvement (or deterioration) data to arrive at a set of mortality rates that can be used to calculate reserves for future events

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

6 copy 2020 Society of Actuaries All rights reserved

Most recent relevant historical MI data (10-year moving average)Historical Data

bull Age- and gender-basedbull Use of a long-term consistent source of population data Social Security Administration (SSA)

Most recent forecast of future improvements over future period (20 years)Forecasted Expectations

bull Age- and gender-based databull Consistent with historical data and projections (Alt II) available from SSA Trusteesrsquo Annual Report

Weighted average of historical data and forecasted expectationsUnsmoothed MI Scale

bull Average of historical data and forecasted expectations

Unsmoothed MI scale with smoothing process appliedSmoothed MI Scale

Current annual update methodology (2013-2019)

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

7

Preliminary 2020 ndash Current MethodologyBefore manual smoothing

-150

-100

-050

000

050

100

150

200

250

300

0 8 16 24 32 40 48 56 64 72 80 88 96 104

112

Male

Male 2020 Male 2019

-150

-100

-050

000

050

100

150

200

250

300

0 7 14 21 28 35 42 49 56 63 70 77 84 91 98 105

112

119

Female

Female 2020 Female 2019

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

8

Considerations for 2020 amp beyondbull Data - we donrsquot have sufficient data to fully understand the impact of the COVID-19-related mortality shock on

the insured population (anecdotal reports from companies indicate they are seeing a smaller shock)

bull MI scale updates - reflecting a shock in 2020 does not seem in line with the goals for the MI scale updates bull Are shock events more appropriately reflected in capital planning rather than reservesbull An effective vaccine may make COVID-19rsquos impact on MI much shorter than the long-term impact arising from the opioid

epidemicbull However COVID-19 may have potential longer-term impacts that may arise from survivor impaired health health

impacts from delays in health care andor testing for early detection of dread disease etcbull Conversely some experts and models indicate the 20202021 COVID-19 shock is mainly a moving forward of deaths that

would have occurred due to other causes andor comorbidities Might that improve future mortality improvement

bull Precedent for other excess mortality events bull First group to consider the impact of a short-term shock event ndash setting a precedent for other future MI scale workbull The current methodology uses a moving average to ldquosmooth outrdquo the impact of any one year or event bull 20082009 influenza season and the effect of the opioid epidemic ndash the methodology was not adjusted for those events

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

9

LMISG 2020 Recommendation

Apply the historical methodology for 2020 consistent with the past scale updates (2013ndash2019)

Implicationsbull There will be no specific impact included for the 2020 scale for the pandemic shock effectbull The 10-year historical average in the 2022 scale update will include a ldquosmoothedrdquo impact

of the shock as part of the usual methodology

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

10

Future Issuesbull Insured vs general population impacts

bull Some evidence that impact on insured population will be lessbull SOA ldquosocioeconomic decilerdquo study will provide some guidance herebull Consideration of consistent framework and changes to the current methodology (ex

averaging periods)

bull Will COVID-19 have a long-term impact on mortality improvement rates and what will the impact be

bull Lower due to survivor impaired health as well as the indirect effect arising from the virus delaying the early condition diagnosis of dread diseases and preventive treatments

bull Higher due to greater application of good hygienic habits (eg social distancing and washing hands) andor higher utilization of other vaccines (such as the annual flu shot)

bull Need to understand the impact in terms of potential effects on future slope and size of MIbull Impact in light of a COVID-19 vaccine availability and effectiveness

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

copy 2020 Society of Actuaries All rights reserved

Questions

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

copy 2020 Society of Actuaries All rights reserved

Appendix

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

13 copy 2020 Society of Actuaries All rights reserved

General PopulationPattern of excess deaths vs general mortality

AE Ratios ndash COVID-19 vs 2015 VBT Actual deaths = COVID-19

Expected deaths = 2015 Unismoke ANB 2015 VBT rates weighted by 72019 estimated US population

Slope of VBT matches COVID-19 female rates better than male rates

Small ratios gtgtgt only 1 cause of death (COD) in numerator all CODs in denominator

0904

0209

2731

4650

464647

46

1103

0506

2541

7580

6857

4657

lt1

5ndash14

25ndash34

45ndash54

65ndash74

85+

1ndash4

15ndash24

35ndash44

55ndash64

75ndash84

All Ages

Age

Grou

p

US Male

US Female

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

14 copy 2020 Society of Actuaries All rights reserved

2013-2019 MI Scale Update Reports2019 Scale httpswwwsoaorgresourcesexperience-studies2019mortality-improvement

2018 Scale httpswwwsoaorgresourcesexperience-studies20182018-mortality-improvement

2017 Scale httpswwwsoaorgresourcesexperience-studies2017mortality-improvement-2017

2016 Scale httpswwwsoaorgresourcesexperience-studies2016research-mortality-improvement-2016

2015 Scale httpswwwsoaorgresourcesexperience-studies2015research-2015-mortality-improvement-rate-year-end

2014 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-improve-rates-ag-38-year-end-2014

2013 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-imp-rates

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

3

Objective of Annual MI Scale UpdatesAddresses VM20 Incorporation of MI Section 9C3g

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

4

Objective of Annual MI Scale UpdatesLevel of Event Covered ndash Reserve vs Capital VM Introduction

Source VM20 ndash guidance provided on risks to be covered by reserves

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

5

Objective for Annual MI Scale UpdatesLMISG Thinking

Our annual update exercise seeks to apply judgment to historical mortality improvement (or deterioration) data to arrive at a set of mortality rates that can be used to calculate reserves for future events

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

6 copy 2020 Society of Actuaries All rights reserved

Most recent relevant historical MI data (10-year moving average)Historical Data

bull Age- and gender-basedbull Use of a long-term consistent source of population data Social Security Administration (SSA)

Most recent forecast of future improvements over future period (20 years)Forecasted Expectations

bull Age- and gender-based databull Consistent with historical data and projections (Alt II) available from SSA Trusteesrsquo Annual Report

Weighted average of historical data and forecasted expectationsUnsmoothed MI Scale

bull Average of historical data and forecasted expectations

Unsmoothed MI scale with smoothing process appliedSmoothed MI Scale

Current annual update methodology (2013-2019)

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

7

Preliminary 2020 ndash Current MethodologyBefore manual smoothing

-150

-100

-050

000

050

100

150

200

250

300

0 8 16 24 32 40 48 56 64 72 80 88 96 104

112

Male

Male 2020 Male 2019

-150

-100

-050

000

050

100

150

200

250

300

0 7 14 21 28 35 42 49 56 63 70 77 84 91 98 105

112

119

Female

Female 2020 Female 2019

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

8

Considerations for 2020 amp beyondbull Data - we donrsquot have sufficient data to fully understand the impact of the COVID-19-related mortality shock on

the insured population (anecdotal reports from companies indicate they are seeing a smaller shock)

bull MI scale updates - reflecting a shock in 2020 does not seem in line with the goals for the MI scale updates bull Are shock events more appropriately reflected in capital planning rather than reservesbull An effective vaccine may make COVID-19rsquos impact on MI much shorter than the long-term impact arising from the opioid

epidemicbull However COVID-19 may have potential longer-term impacts that may arise from survivor impaired health health

impacts from delays in health care andor testing for early detection of dread disease etcbull Conversely some experts and models indicate the 20202021 COVID-19 shock is mainly a moving forward of deaths that

would have occurred due to other causes andor comorbidities Might that improve future mortality improvement

bull Precedent for other excess mortality events bull First group to consider the impact of a short-term shock event ndash setting a precedent for other future MI scale workbull The current methodology uses a moving average to ldquosmooth outrdquo the impact of any one year or event bull 20082009 influenza season and the effect of the opioid epidemic ndash the methodology was not adjusted for those events

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

9

LMISG 2020 Recommendation

Apply the historical methodology for 2020 consistent with the past scale updates (2013ndash2019)

Implicationsbull There will be no specific impact included for the 2020 scale for the pandemic shock effectbull The 10-year historical average in the 2022 scale update will include a ldquosmoothedrdquo impact

of the shock as part of the usual methodology

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

10

Future Issuesbull Insured vs general population impacts

bull Some evidence that impact on insured population will be lessbull SOA ldquosocioeconomic decilerdquo study will provide some guidance herebull Consideration of consistent framework and changes to the current methodology (ex

averaging periods)

bull Will COVID-19 have a long-term impact on mortality improvement rates and what will the impact be

bull Lower due to survivor impaired health as well as the indirect effect arising from the virus delaying the early condition diagnosis of dread diseases and preventive treatments

bull Higher due to greater application of good hygienic habits (eg social distancing and washing hands) andor higher utilization of other vaccines (such as the annual flu shot)

bull Need to understand the impact in terms of potential effects on future slope and size of MIbull Impact in light of a COVID-19 vaccine availability and effectiveness

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

copy 2020 Society of Actuaries All rights reserved

Questions

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

copy 2020 Society of Actuaries All rights reserved

Appendix

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

13 copy 2020 Society of Actuaries All rights reserved

General PopulationPattern of excess deaths vs general mortality

AE Ratios ndash COVID-19 vs 2015 VBT Actual deaths = COVID-19

Expected deaths = 2015 Unismoke ANB 2015 VBT rates weighted by 72019 estimated US population

Slope of VBT matches COVID-19 female rates better than male rates

Small ratios gtgtgt only 1 cause of death (COD) in numerator all CODs in denominator

0904

0209

2731

4650

464647

46

1103

0506

2541

7580

6857

4657

lt1

5ndash14

25ndash34

45ndash54

65ndash74

85+

1ndash4

15ndash24

35ndash44

55ndash64

75ndash84

All Ages

Age

Grou

p

US Male

US Female

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

14 copy 2020 Society of Actuaries All rights reserved

2013-2019 MI Scale Update Reports2019 Scale httpswwwsoaorgresourcesexperience-studies2019mortality-improvement

2018 Scale httpswwwsoaorgresourcesexperience-studies20182018-mortality-improvement

2017 Scale httpswwwsoaorgresourcesexperience-studies2017mortality-improvement-2017

2016 Scale httpswwwsoaorgresourcesexperience-studies2016research-mortality-improvement-2016

2015 Scale httpswwwsoaorgresourcesexperience-studies2015research-2015-mortality-improvement-rate-year-end

2014 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-improve-rates-ag-38-year-end-2014

2013 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-imp-rates

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

4

Objective of Annual MI Scale UpdatesLevel of Event Covered ndash Reserve vs Capital VM Introduction

Source VM20 ndash guidance provided on risks to be covered by reserves

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

5

Objective for Annual MI Scale UpdatesLMISG Thinking

Our annual update exercise seeks to apply judgment to historical mortality improvement (or deterioration) data to arrive at a set of mortality rates that can be used to calculate reserves for future events

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

6 copy 2020 Society of Actuaries All rights reserved

Most recent relevant historical MI data (10-year moving average)Historical Data

bull Age- and gender-basedbull Use of a long-term consistent source of population data Social Security Administration (SSA)

Most recent forecast of future improvements over future period (20 years)Forecasted Expectations

bull Age- and gender-based databull Consistent with historical data and projections (Alt II) available from SSA Trusteesrsquo Annual Report

Weighted average of historical data and forecasted expectationsUnsmoothed MI Scale

bull Average of historical data and forecasted expectations

Unsmoothed MI scale with smoothing process appliedSmoothed MI Scale

Current annual update methodology (2013-2019)

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

7

Preliminary 2020 ndash Current MethodologyBefore manual smoothing

-150

-100

-050

000

050

100

150

200

250

300

0 8 16 24 32 40 48 56 64 72 80 88 96 104

112

Male

Male 2020 Male 2019

-150

-100

-050

000

050

100

150

200

250

300

0 7 14 21 28 35 42 49 56 63 70 77 84 91 98 105

112

119

Female

Female 2020 Female 2019

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

8

Considerations for 2020 amp beyondbull Data - we donrsquot have sufficient data to fully understand the impact of the COVID-19-related mortality shock on

the insured population (anecdotal reports from companies indicate they are seeing a smaller shock)

bull MI scale updates - reflecting a shock in 2020 does not seem in line with the goals for the MI scale updates bull Are shock events more appropriately reflected in capital planning rather than reservesbull An effective vaccine may make COVID-19rsquos impact on MI much shorter than the long-term impact arising from the opioid

epidemicbull However COVID-19 may have potential longer-term impacts that may arise from survivor impaired health health

impacts from delays in health care andor testing for early detection of dread disease etcbull Conversely some experts and models indicate the 20202021 COVID-19 shock is mainly a moving forward of deaths that

would have occurred due to other causes andor comorbidities Might that improve future mortality improvement

bull Precedent for other excess mortality events bull First group to consider the impact of a short-term shock event ndash setting a precedent for other future MI scale workbull The current methodology uses a moving average to ldquosmooth outrdquo the impact of any one year or event bull 20082009 influenza season and the effect of the opioid epidemic ndash the methodology was not adjusted for those events

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

9

LMISG 2020 Recommendation

Apply the historical methodology for 2020 consistent with the past scale updates (2013ndash2019)

Implicationsbull There will be no specific impact included for the 2020 scale for the pandemic shock effectbull The 10-year historical average in the 2022 scale update will include a ldquosmoothedrdquo impact

of the shock as part of the usual methodology

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

10

Future Issuesbull Insured vs general population impacts

bull Some evidence that impact on insured population will be lessbull SOA ldquosocioeconomic decilerdquo study will provide some guidance herebull Consideration of consistent framework and changes to the current methodology (ex

averaging periods)

bull Will COVID-19 have a long-term impact on mortality improvement rates and what will the impact be

bull Lower due to survivor impaired health as well as the indirect effect arising from the virus delaying the early condition diagnosis of dread diseases and preventive treatments

bull Higher due to greater application of good hygienic habits (eg social distancing and washing hands) andor higher utilization of other vaccines (such as the annual flu shot)

bull Need to understand the impact in terms of potential effects on future slope and size of MIbull Impact in light of a COVID-19 vaccine availability and effectiveness

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

copy 2020 Society of Actuaries All rights reserved

Questions

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

copy 2020 Society of Actuaries All rights reserved

Appendix

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

13 copy 2020 Society of Actuaries All rights reserved

General PopulationPattern of excess deaths vs general mortality

AE Ratios ndash COVID-19 vs 2015 VBT Actual deaths = COVID-19

Expected deaths = 2015 Unismoke ANB 2015 VBT rates weighted by 72019 estimated US population

Slope of VBT matches COVID-19 female rates better than male rates

Small ratios gtgtgt only 1 cause of death (COD) in numerator all CODs in denominator

0904

0209

2731

4650

464647

46

1103

0506

2541

7580

6857

4657

lt1

5ndash14

25ndash34

45ndash54

65ndash74

85+

1ndash4

15ndash24

35ndash44

55ndash64

75ndash84

All Ages

Age

Grou

p

US Male

US Female

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

14 copy 2020 Society of Actuaries All rights reserved

2013-2019 MI Scale Update Reports2019 Scale httpswwwsoaorgresourcesexperience-studies2019mortality-improvement

2018 Scale httpswwwsoaorgresourcesexperience-studies20182018-mortality-improvement

2017 Scale httpswwwsoaorgresourcesexperience-studies2017mortality-improvement-2017

2016 Scale httpswwwsoaorgresourcesexperience-studies2016research-mortality-improvement-2016

2015 Scale httpswwwsoaorgresourcesexperience-studies2015research-2015-mortality-improvement-rate-year-end

2014 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-improve-rates-ag-38-year-end-2014

2013 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-imp-rates

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

5

Objective for Annual MI Scale UpdatesLMISG Thinking

Our annual update exercise seeks to apply judgment to historical mortality improvement (or deterioration) data to arrive at a set of mortality rates that can be used to calculate reserves for future events

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

6 copy 2020 Society of Actuaries All rights reserved

Most recent relevant historical MI data (10-year moving average)Historical Data

bull Age- and gender-basedbull Use of a long-term consistent source of population data Social Security Administration (SSA)

Most recent forecast of future improvements over future period (20 years)Forecasted Expectations

bull Age- and gender-based databull Consistent with historical data and projections (Alt II) available from SSA Trusteesrsquo Annual Report

Weighted average of historical data and forecasted expectationsUnsmoothed MI Scale

bull Average of historical data and forecasted expectations

Unsmoothed MI scale with smoothing process appliedSmoothed MI Scale

Current annual update methodology (2013-2019)

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

7

Preliminary 2020 ndash Current MethodologyBefore manual smoothing

-150

-100

-050

000

050

100

150

200

250

300

0 8 16 24 32 40 48 56 64 72 80 88 96 104

112

Male

Male 2020 Male 2019

-150

-100

-050

000

050

100

150

200

250

300

0 7 14 21 28 35 42 49 56 63 70 77 84 91 98 105

112

119

Female

Female 2020 Female 2019

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

8

Considerations for 2020 amp beyondbull Data - we donrsquot have sufficient data to fully understand the impact of the COVID-19-related mortality shock on

the insured population (anecdotal reports from companies indicate they are seeing a smaller shock)

bull MI scale updates - reflecting a shock in 2020 does not seem in line with the goals for the MI scale updates bull Are shock events more appropriately reflected in capital planning rather than reservesbull An effective vaccine may make COVID-19rsquos impact on MI much shorter than the long-term impact arising from the opioid

epidemicbull However COVID-19 may have potential longer-term impacts that may arise from survivor impaired health health

impacts from delays in health care andor testing for early detection of dread disease etcbull Conversely some experts and models indicate the 20202021 COVID-19 shock is mainly a moving forward of deaths that

would have occurred due to other causes andor comorbidities Might that improve future mortality improvement

bull Precedent for other excess mortality events bull First group to consider the impact of a short-term shock event ndash setting a precedent for other future MI scale workbull The current methodology uses a moving average to ldquosmooth outrdquo the impact of any one year or event bull 20082009 influenza season and the effect of the opioid epidemic ndash the methodology was not adjusted for those events

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

9

LMISG 2020 Recommendation

Apply the historical methodology for 2020 consistent with the past scale updates (2013ndash2019)

Implicationsbull There will be no specific impact included for the 2020 scale for the pandemic shock effectbull The 10-year historical average in the 2022 scale update will include a ldquosmoothedrdquo impact

of the shock as part of the usual methodology

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

10

Future Issuesbull Insured vs general population impacts

bull Some evidence that impact on insured population will be lessbull SOA ldquosocioeconomic decilerdquo study will provide some guidance herebull Consideration of consistent framework and changes to the current methodology (ex

averaging periods)

bull Will COVID-19 have a long-term impact on mortality improvement rates and what will the impact be

bull Lower due to survivor impaired health as well as the indirect effect arising from the virus delaying the early condition diagnosis of dread diseases and preventive treatments

bull Higher due to greater application of good hygienic habits (eg social distancing and washing hands) andor higher utilization of other vaccines (such as the annual flu shot)

bull Need to understand the impact in terms of potential effects on future slope and size of MIbull Impact in light of a COVID-19 vaccine availability and effectiveness

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

copy 2020 Society of Actuaries All rights reserved

Questions

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

copy 2020 Society of Actuaries All rights reserved

Appendix

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

13 copy 2020 Society of Actuaries All rights reserved

General PopulationPattern of excess deaths vs general mortality

AE Ratios ndash COVID-19 vs 2015 VBT Actual deaths = COVID-19

Expected deaths = 2015 Unismoke ANB 2015 VBT rates weighted by 72019 estimated US population

Slope of VBT matches COVID-19 female rates better than male rates

Small ratios gtgtgt only 1 cause of death (COD) in numerator all CODs in denominator

0904

0209

2731

4650

464647

46

1103

0506

2541

7580

6857

4657

lt1

5ndash14

25ndash34

45ndash54

65ndash74

85+

1ndash4

15ndash24

35ndash44

55ndash64

75ndash84

All Ages

Age

Grou

p

US Male

US Female

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

14 copy 2020 Society of Actuaries All rights reserved

2013-2019 MI Scale Update Reports2019 Scale httpswwwsoaorgresourcesexperience-studies2019mortality-improvement

2018 Scale httpswwwsoaorgresourcesexperience-studies20182018-mortality-improvement

2017 Scale httpswwwsoaorgresourcesexperience-studies2017mortality-improvement-2017

2016 Scale httpswwwsoaorgresourcesexperience-studies2016research-mortality-improvement-2016

2015 Scale httpswwwsoaorgresourcesexperience-studies2015research-2015-mortality-improvement-rate-year-end

2014 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-improve-rates-ag-38-year-end-2014

2013 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-imp-rates

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

6 copy 2020 Society of Actuaries All rights reserved

Most recent relevant historical MI data (10-year moving average)Historical Data

bull Age- and gender-basedbull Use of a long-term consistent source of population data Social Security Administration (SSA)

Most recent forecast of future improvements over future period (20 years)Forecasted Expectations

bull Age- and gender-based databull Consistent with historical data and projections (Alt II) available from SSA Trusteesrsquo Annual Report

Weighted average of historical data and forecasted expectationsUnsmoothed MI Scale

bull Average of historical data and forecasted expectations

Unsmoothed MI scale with smoothing process appliedSmoothed MI Scale

Current annual update methodology (2013-2019)

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

7

Preliminary 2020 ndash Current MethodologyBefore manual smoothing

-150

-100

-050

000

050

100

150

200

250

300

0 8 16 24 32 40 48 56 64 72 80 88 96 104

112

Male

Male 2020 Male 2019

-150

-100

-050

000

050

100

150

200

250

300

0 7 14 21 28 35 42 49 56 63 70 77 84 91 98 105

112

119

Female

Female 2020 Female 2019

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

8

Considerations for 2020 amp beyondbull Data - we donrsquot have sufficient data to fully understand the impact of the COVID-19-related mortality shock on

the insured population (anecdotal reports from companies indicate they are seeing a smaller shock)

bull MI scale updates - reflecting a shock in 2020 does not seem in line with the goals for the MI scale updates bull Are shock events more appropriately reflected in capital planning rather than reservesbull An effective vaccine may make COVID-19rsquos impact on MI much shorter than the long-term impact arising from the opioid

epidemicbull However COVID-19 may have potential longer-term impacts that may arise from survivor impaired health health

impacts from delays in health care andor testing for early detection of dread disease etcbull Conversely some experts and models indicate the 20202021 COVID-19 shock is mainly a moving forward of deaths that

would have occurred due to other causes andor comorbidities Might that improve future mortality improvement

bull Precedent for other excess mortality events bull First group to consider the impact of a short-term shock event ndash setting a precedent for other future MI scale workbull The current methodology uses a moving average to ldquosmooth outrdquo the impact of any one year or event bull 20082009 influenza season and the effect of the opioid epidemic ndash the methodology was not adjusted for those events

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

9

LMISG 2020 Recommendation

Apply the historical methodology for 2020 consistent with the past scale updates (2013ndash2019)

Implicationsbull There will be no specific impact included for the 2020 scale for the pandemic shock effectbull The 10-year historical average in the 2022 scale update will include a ldquosmoothedrdquo impact

of the shock as part of the usual methodology

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

10

Future Issuesbull Insured vs general population impacts

bull Some evidence that impact on insured population will be lessbull SOA ldquosocioeconomic decilerdquo study will provide some guidance herebull Consideration of consistent framework and changes to the current methodology (ex

averaging periods)

bull Will COVID-19 have a long-term impact on mortality improvement rates and what will the impact be

bull Lower due to survivor impaired health as well as the indirect effect arising from the virus delaying the early condition diagnosis of dread diseases and preventive treatments

bull Higher due to greater application of good hygienic habits (eg social distancing and washing hands) andor higher utilization of other vaccines (such as the annual flu shot)

bull Need to understand the impact in terms of potential effects on future slope and size of MIbull Impact in light of a COVID-19 vaccine availability and effectiveness

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

copy 2020 Society of Actuaries All rights reserved

Questions

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

copy 2020 Society of Actuaries All rights reserved

Appendix

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

13 copy 2020 Society of Actuaries All rights reserved

General PopulationPattern of excess deaths vs general mortality

AE Ratios ndash COVID-19 vs 2015 VBT Actual deaths = COVID-19

Expected deaths = 2015 Unismoke ANB 2015 VBT rates weighted by 72019 estimated US population

Slope of VBT matches COVID-19 female rates better than male rates

Small ratios gtgtgt only 1 cause of death (COD) in numerator all CODs in denominator

0904

0209

2731

4650

464647

46

1103

0506

2541

7580

6857

4657

lt1

5ndash14

25ndash34

45ndash54

65ndash74

85+

1ndash4

15ndash24

35ndash44

55ndash64

75ndash84

All Ages

Age

Grou

p

US Male

US Female

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

14 copy 2020 Society of Actuaries All rights reserved

2013-2019 MI Scale Update Reports2019 Scale httpswwwsoaorgresourcesexperience-studies2019mortality-improvement

2018 Scale httpswwwsoaorgresourcesexperience-studies20182018-mortality-improvement

2017 Scale httpswwwsoaorgresourcesexperience-studies2017mortality-improvement-2017

2016 Scale httpswwwsoaorgresourcesexperience-studies2016research-mortality-improvement-2016

2015 Scale httpswwwsoaorgresourcesexperience-studies2015research-2015-mortality-improvement-rate-year-end

2014 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-improve-rates-ag-38-year-end-2014

2013 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-imp-rates

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

7

Preliminary 2020 ndash Current MethodologyBefore manual smoothing

-150

-100

-050

000

050

100

150

200

250

300

0 8 16 24 32 40 48 56 64 72 80 88 96 104

112

Male

Male 2020 Male 2019

-150

-100

-050

000

050

100

150

200

250

300

0 7 14 21 28 35 42 49 56 63 70 77 84 91 98 105

112

119

Female

Female 2020 Female 2019

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

8

Considerations for 2020 amp beyondbull Data - we donrsquot have sufficient data to fully understand the impact of the COVID-19-related mortality shock on

the insured population (anecdotal reports from companies indicate they are seeing a smaller shock)

bull MI scale updates - reflecting a shock in 2020 does not seem in line with the goals for the MI scale updates bull Are shock events more appropriately reflected in capital planning rather than reservesbull An effective vaccine may make COVID-19rsquos impact on MI much shorter than the long-term impact arising from the opioid

epidemicbull However COVID-19 may have potential longer-term impacts that may arise from survivor impaired health health

impacts from delays in health care andor testing for early detection of dread disease etcbull Conversely some experts and models indicate the 20202021 COVID-19 shock is mainly a moving forward of deaths that

would have occurred due to other causes andor comorbidities Might that improve future mortality improvement

bull Precedent for other excess mortality events bull First group to consider the impact of a short-term shock event ndash setting a precedent for other future MI scale workbull The current methodology uses a moving average to ldquosmooth outrdquo the impact of any one year or event bull 20082009 influenza season and the effect of the opioid epidemic ndash the methodology was not adjusted for those events

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

9

LMISG 2020 Recommendation

Apply the historical methodology for 2020 consistent with the past scale updates (2013ndash2019)

Implicationsbull There will be no specific impact included for the 2020 scale for the pandemic shock effectbull The 10-year historical average in the 2022 scale update will include a ldquosmoothedrdquo impact

of the shock as part of the usual methodology

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

10

Future Issuesbull Insured vs general population impacts

bull Some evidence that impact on insured population will be lessbull SOA ldquosocioeconomic decilerdquo study will provide some guidance herebull Consideration of consistent framework and changes to the current methodology (ex

averaging periods)

bull Will COVID-19 have a long-term impact on mortality improvement rates and what will the impact be

bull Lower due to survivor impaired health as well as the indirect effect arising from the virus delaying the early condition diagnosis of dread diseases and preventive treatments

bull Higher due to greater application of good hygienic habits (eg social distancing and washing hands) andor higher utilization of other vaccines (such as the annual flu shot)

bull Need to understand the impact in terms of potential effects on future slope and size of MIbull Impact in light of a COVID-19 vaccine availability and effectiveness

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

copy 2020 Society of Actuaries All rights reserved

Questions

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

copy 2020 Society of Actuaries All rights reserved

Appendix

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

13 copy 2020 Society of Actuaries All rights reserved

General PopulationPattern of excess deaths vs general mortality

AE Ratios ndash COVID-19 vs 2015 VBT Actual deaths = COVID-19

Expected deaths = 2015 Unismoke ANB 2015 VBT rates weighted by 72019 estimated US population

Slope of VBT matches COVID-19 female rates better than male rates

Small ratios gtgtgt only 1 cause of death (COD) in numerator all CODs in denominator

0904

0209

2731

4650

464647

46

1103

0506

2541

7580

6857

4657

lt1

5ndash14

25ndash34

45ndash54

65ndash74

85+

1ndash4

15ndash24

35ndash44

55ndash64

75ndash84

All Ages

Age

Grou

p

US Male

US Female

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

14 copy 2020 Society of Actuaries All rights reserved

2013-2019 MI Scale Update Reports2019 Scale httpswwwsoaorgresourcesexperience-studies2019mortality-improvement

2018 Scale httpswwwsoaorgresourcesexperience-studies20182018-mortality-improvement

2017 Scale httpswwwsoaorgresourcesexperience-studies2017mortality-improvement-2017

2016 Scale httpswwwsoaorgresourcesexperience-studies2016research-mortality-improvement-2016

2015 Scale httpswwwsoaorgresourcesexperience-studies2015research-2015-mortality-improvement-rate-year-end

2014 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-improve-rates-ag-38-year-end-2014

2013 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-imp-rates

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

8

Considerations for 2020 amp beyondbull Data - we donrsquot have sufficient data to fully understand the impact of the COVID-19-related mortality shock on

the insured population (anecdotal reports from companies indicate they are seeing a smaller shock)

bull MI scale updates - reflecting a shock in 2020 does not seem in line with the goals for the MI scale updates bull Are shock events more appropriately reflected in capital planning rather than reservesbull An effective vaccine may make COVID-19rsquos impact on MI much shorter than the long-term impact arising from the opioid

epidemicbull However COVID-19 may have potential longer-term impacts that may arise from survivor impaired health health

impacts from delays in health care andor testing for early detection of dread disease etcbull Conversely some experts and models indicate the 20202021 COVID-19 shock is mainly a moving forward of deaths that

would have occurred due to other causes andor comorbidities Might that improve future mortality improvement

bull Precedent for other excess mortality events bull First group to consider the impact of a short-term shock event ndash setting a precedent for other future MI scale workbull The current methodology uses a moving average to ldquosmooth outrdquo the impact of any one year or event bull 20082009 influenza season and the effect of the opioid epidemic ndash the methodology was not adjusted for those events

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

9

LMISG 2020 Recommendation

Apply the historical methodology for 2020 consistent with the past scale updates (2013ndash2019)

Implicationsbull There will be no specific impact included for the 2020 scale for the pandemic shock effectbull The 10-year historical average in the 2022 scale update will include a ldquosmoothedrdquo impact

of the shock as part of the usual methodology

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

10

Future Issuesbull Insured vs general population impacts

bull Some evidence that impact on insured population will be lessbull SOA ldquosocioeconomic decilerdquo study will provide some guidance herebull Consideration of consistent framework and changes to the current methodology (ex

averaging periods)

bull Will COVID-19 have a long-term impact on mortality improvement rates and what will the impact be

bull Lower due to survivor impaired health as well as the indirect effect arising from the virus delaying the early condition diagnosis of dread diseases and preventive treatments

bull Higher due to greater application of good hygienic habits (eg social distancing and washing hands) andor higher utilization of other vaccines (such as the annual flu shot)

bull Need to understand the impact in terms of potential effects on future slope and size of MIbull Impact in light of a COVID-19 vaccine availability and effectiveness

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

copy 2020 Society of Actuaries All rights reserved

Questions

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

copy 2020 Society of Actuaries All rights reserved

Appendix

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

13 copy 2020 Society of Actuaries All rights reserved

General PopulationPattern of excess deaths vs general mortality

AE Ratios ndash COVID-19 vs 2015 VBT Actual deaths = COVID-19

Expected deaths = 2015 Unismoke ANB 2015 VBT rates weighted by 72019 estimated US population

Slope of VBT matches COVID-19 female rates better than male rates

Small ratios gtgtgt only 1 cause of death (COD) in numerator all CODs in denominator

0904

0209

2731

4650

464647

46

1103

0506

2541

7580

6857

4657

lt1

5ndash14

25ndash34

45ndash54

65ndash74

85+

1ndash4

15ndash24

35ndash44

55ndash64

75ndash84

All Ages

Age

Grou

p

US Male

US Female

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

14 copy 2020 Society of Actuaries All rights reserved

2013-2019 MI Scale Update Reports2019 Scale httpswwwsoaorgresourcesexperience-studies2019mortality-improvement

2018 Scale httpswwwsoaorgresourcesexperience-studies20182018-mortality-improvement

2017 Scale httpswwwsoaorgresourcesexperience-studies2017mortality-improvement-2017

2016 Scale httpswwwsoaorgresourcesexperience-studies2016research-mortality-improvement-2016

2015 Scale httpswwwsoaorgresourcesexperience-studies2015research-2015-mortality-improvement-rate-year-end

2014 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-improve-rates-ag-38-year-end-2014

2013 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-imp-rates

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

9

LMISG 2020 Recommendation

Apply the historical methodology for 2020 consistent with the past scale updates (2013ndash2019)

Implicationsbull There will be no specific impact included for the 2020 scale for the pandemic shock effectbull The 10-year historical average in the 2022 scale update will include a ldquosmoothedrdquo impact

of the shock as part of the usual methodology

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

10

Future Issuesbull Insured vs general population impacts

bull Some evidence that impact on insured population will be lessbull SOA ldquosocioeconomic decilerdquo study will provide some guidance herebull Consideration of consistent framework and changes to the current methodology (ex

averaging periods)

bull Will COVID-19 have a long-term impact on mortality improvement rates and what will the impact be

bull Lower due to survivor impaired health as well as the indirect effect arising from the virus delaying the early condition diagnosis of dread diseases and preventive treatments

bull Higher due to greater application of good hygienic habits (eg social distancing and washing hands) andor higher utilization of other vaccines (such as the annual flu shot)

bull Need to understand the impact in terms of potential effects on future slope and size of MIbull Impact in light of a COVID-19 vaccine availability and effectiveness

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

copy 2020 Society of Actuaries All rights reserved

Questions

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

copy 2020 Society of Actuaries All rights reserved

Appendix

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

13 copy 2020 Society of Actuaries All rights reserved

General PopulationPattern of excess deaths vs general mortality

AE Ratios ndash COVID-19 vs 2015 VBT Actual deaths = COVID-19

Expected deaths = 2015 Unismoke ANB 2015 VBT rates weighted by 72019 estimated US population

Slope of VBT matches COVID-19 female rates better than male rates

Small ratios gtgtgt only 1 cause of death (COD) in numerator all CODs in denominator

0904

0209

2731

4650

464647

46

1103

0506

2541

7580

6857

4657

lt1

5ndash14

25ndash34

45ndash54

65ndash74

85+

1ndash4

15ndash24

35ndash44

55ndash64

75ndash84

All Ages

Age

Grou

p

US Male

US Female

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

14 copy 2020 Society of Actuaries All rights reserved

2013-2019 MI Scale Update Reports2019 Scale httpswwwsoaorgresourcesexperience-studies2019mortality-improvement

2018 Scale httpswwwsoaorgresourcesexperience-studies20182018-mortality-improvement

2017 Scale httpswwwsoaorgresourcesexperience-studies2017mortality-improvement-2017

2016 Scale httpswwwsoaorgresourcesexperience-studies2016research-mortality-improvement-2016

2015 Scale httpswwwsoaorgresourcesexperience-studies2015research-2015-mortality-improvement-rate-year-end

2014 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-improve-rates-ag-38-year-end-2014

2013 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-imp-rates

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

10

Future Issuesbull Insured vs general population impacts

bull Some evidence that impact on insured population will be lessbull SOA ldquosocioeconomic decilerdquo study will provide some guidance herebull Consideration of consistent framework and changes to the current methodology (ex

averaging periods)

bull Will COVID-19 have a long-term impact on mortality improvement rates and what will the impact be

bull Lower due to survivor impaired health as well as the indirect effect arising from the virus delaying the early condition diagnosis of dread diseases and preventive treatments

bull Higher due to greater application of good hygienic habits (eg social distancing and washing hands) andor higher utilization of other vaccines (such as the annual flu shot)

bull Need to understand the impact in terms of potential effects on future slope and size of MIbull Impact in light of a COVID-19 vaccine availability and effectiveness

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

copy 2020 Society of Actuaries All rights reserved

Questions

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

copy 2020 Society of Actuaries All rights reserved

Appendix

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

13 copy 2020 Society of Actuaries All rights reserved

General PopulationPattern of excess deaths vs general mortality

AE Ratios ndash COVID-19 vs 2015 VBT Actual deaths = COVID-19

Expected deaths = 2015 Unismoke ANB 2015 VBT rates weighted by 72019 estimated US population

Slope of VBT matches COVID-19 female rates better than male rates

Small ratios gtgtgt only 1 cause of death (COD) in numerator all CODs in denominator

0904

0209

2731

4650

464647

46

1103

0506

2541

7580

6857

4657

lt1

5ndash14

25ndash34

45ndash54

65ndash74

85+

1ndash4

15ndash24

35ndash44

55ndash64

75ndash84

All Ages

Age

Grou

p

US Male

US Female

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

14 copy 2020 Society of Actuaries All rights reserved

2013-2019 MI Scale Update Reports2019 Scale httpswwwsoaorgresourcesexperience-studies2019mortality-improvement

2018 Scale httpswwwsoaorgresourcesexperience-studies20182018-mortality-improvement

2017 Scale httpswwwsoaorgresourcesexperience-studies2017mortality-improvement-2017

2016 Scale httpswwwsoaorgresourcesexperience-studies2016research-mortality-improvement-2016

2015 Scale httpswwwsoaorgresourcesexperience-studies2015research-2015-mortality-improvement-rate-year-end

2014 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-improve-rates-ag-38-year-end-2014

2013 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-imp-rates

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

copy 2020 Society of Actuaries All rights reserved

Questions

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

copy 2020 Society of Actuaries All rights reserved

Appendix

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

13 copy 2020 Society of Actuaries All rights reserved

General PopulationPattern of excess deaths vs general mortality

AE Ratios ndash COVID-19 vs 2015 VBT Actual deaths = COVID-19

Expected deaths = 2015 Unismoke ANB 2015 VBT rates weighted by 72019 estimated US population

Slope of VBT matches COVID-19 female rates better than male rates

Small ratios gtgtgt only 1 cause of death (COD) in numerator all CODs in denominator

0904

0209

2731

4650

464647

46

1103

0506

2541

7580

6857

4657

lt1

5ndash14

25ndash34

45ndash54

65ndash74

85+

1ndash4

15ndash24

35ndash44

55ndash64

75ndash84

All Ages

Age

Grou

p

US Male

US Female

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

14 copy 2020 Society of Actuaries All rights reserved

2013-2019 MI Scale Update Reports2019 Scale httpswwwsoaorgresourcesexperience-studies2019mortality-improvement

2018 Scale httpswwwsoaorgresourcesexperience-studies20182018-mortality-improvement

2017 Scale httpswwwsoaorgresourcesexperience-studies2017mortality-improvement-2017

2016 Scale httpswwwsoaorgresourcesexperience-studies2016research-mortality-improvement-2016

2015 Scale httpswwwsoaorgresourcesexperience-studies2015research-2015-mortality-improvement-rate-year-end

2014 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-improve-rates-ag-38-year-end-2014

2013 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-imp-rates

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

copy 2020 Society of Actuaries All rights reserved

Appendix

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

13 copy 2020 Society of Actuaries All rights reserved

General PopulationPattern of excess deaths vs general mortality

AE Ratios ndash COVID-19 vs 2015 VBT Actual deaths = COVID-19

Expected deaths = 2015 Unismoke ANB 2015 VBT rates weighted by 72019 estimated US population

Slope of VBT matches COVID-19 female rates better than male rates

Small ratios gtgtgt only 1 cause of death (COD) in numerator all CODs in denominator

0904

0209

2731

4650

464647

46

1103

0506

2541

7580

6857

4657

lt1

5ndash14

25ndash34

45ndash54

65ndash74

85+

1ndash4

15ndash24

35ndash44

55ndash64

75ndash84

All Ages

Age

Grou

p

US Male

US Female

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

14 copy 2020 Society of Actuaries All rights reserved

2013-2019 MI Scale Update Reports2019 Scale httpswwwsoaorgresourcesexperience-studies2019mortality-improvement

2018 Scale httpswwwsoaorgresourcesexperience-studies20182018-mortality-improvement

2017 Scale httpswwwsoaorgresourcesexperience-studies2017mortality-improvement-2017

2016 Scale httpswwwsoaorgresourcesexperience-studies2016research-mortality-improvement-2016

2015 Scale httpswwwsoaorgresourcesexperience-studies2015research-2015-mortality-improvement-rate-year-end

2014 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-improve-rates-ag-38-year-end-2014

2013 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-imp-rates

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

13 copy 2020 Society of Actuaries All rights reserved

General PopulationPattern of excess deaths vs general mortality

AE Ratios ndash COVID-19 vs 2015 VBT Actual deaths = COVID-19

Expected deaths = 2015 Unismoke ANB 2015 VBT rates weighted by 72019 estimated US population

Slope of VBT matches COVID-19 female rates better than male rates

Small ratios gtgtgt only 1 cause of death (COD) in numerator all CODs in denominator

0904

0209

2731

4650

464647

46

1103

0506

2541

7580

6857

4657

lt1

5ndash14

25ndash34

45ndash54

65ndash74

85+

1ndash4

15ndash24

35ndash44

55ndash64

75ndash84

All Ages

Age

Grou

p

US Male

US Female

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

14 copy 2020 Society of Actuaries All rights reserved

2013-2019 MI Scale Update Reports2019 Scale httpswwwsoaorgresourcesexperience-studies2019mortality-improvement

2018 Scale httpswwwsoaorgresourcesexperience-studies20182018-mortality-improvement

2017 Scale httpswwwsoaorgresourcesexperience-studies2017mortality-improvement-2017

2016 Scale httpswwwsoaorgresourcesexperience-studies2016research-mortality-improvement-2016

2015 Scale httpswwwsoaorgresourcesexperience-studies2015research-2015-mortality-improvement-rate-year-end

2014 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-improve-rates-ag-38-year-end-2014

2013 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-imp-rates

copy 2020 American Academy of Actuaries All rights reservedMay not be reproduced without express permission

14 copy 2020 Society of Actuaries All rights reserved

2013-2019 MI Scale Update Reports2019 Scale httpswwwsoaorgresourcesexperience-studies2019mortality-improvement

2018 Scale httpswwwsoaorgresourcesexperience-studies20182018-mortality-improvement

2017 Scale httpswwwsoaorgresourcesexperience-studies2017mortality-improvement-2017

2016 Scale httpswwwsoaorgresourcesexperience-studies2016research-mortality-improvement-2016

2015 Scale httpswwwsoaorgresourcesexperience-studies2015research-2015-mortality-improvement-rate-year-end

2014 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-improve-rates-ag-38-year-end-2014

2013 Scale httpswwwsoaorgresourcesexperience-studies2014research-2014-mort-imp-rates