Mortality Improvement Trends and Assumption Setting

Transcription

1 Mortality Improvement Trends and Assumption Setting Marianne Purushotham, FSA, MAAA SEAC Annual Meeting November 15, 2012 Topics to be covered Review of historical mortality improvement trends US population mortality trend differentials Setting assumptions regarding the future current practices 2 2 1

2 Review of Historical Mortality Improvement Trends Mortality Improvement Global Trends in Life Expectancy at Birth ( ) United States Canada England/Wales China Japan India Italy Germany Greece Females 1900 F 1950 F 1990 F 2009 F United States Canada England/Wales China Japan India Italy Germany Greece Source: World Health Organization, WHO Males 1900 M 1950 M 1990 M 2009 M 4 2

3 Mortality Improvement Trends by Attained Age Ages Under % 120% Vietnam War Male Mortality Percent of 1950 Levels AIDS impact 100% 80% 60% 40% 20% 0% Source: Human Mortality Database, HMD Mortality Improvement Trends by Attained Age Ages Under % Female Mortality Percent of 1950 Levels 100% 80% 60% 40% 20% 0% Source: Human Mortality Database, HMD 3

4 Mortality Improvement Trends by Attained Age Ages 50 and older 140% Male Mortality Percent of 1950 Levels 120% 100% 80% 60% 40% 20% 0% Source: Human Mortality Database, HMD Mortality Improvement Trends by Attained Age Ages 50 and older 140% Female Mortality Percent of 1950 Levels 120% 100% 80% 60% 40% 20% 0% Source: Human Mortality Database, HMD 4

5 Mortality Improvement - Males Heat Maps 2 dimensional view of MI data SSA data, Source: Social Security Administration, SOA 9 Mortality Improvement - Females Heat Maps 2 dimensional view of MI data SSA data, Source: Social Security Administration, SOA 10 5

6 Mortality Improvement - Males England & Wales Canada US Source: Social Security Administration, SOA RPEC Committee 11 Mortality Improvement - Females England & Wales Canada US Source: Social Security Administration, SOA RPEC Committee 12 6

7 US Population - Mortality Improvement Trend Differentials Sources of Variation in Population Longevity Gender Men injuries, alcohol, smoking Females smaller, resistance to infection/disease Region/Geography Diet Behaviors Climate Race Socioeconomic Status Education, Income/Wealth Other Marital status Public Health Advances 14 Access to Medical Care and their interrelationships 7

8 U.S. Population Mortality Improvements - Gender 4.0% 3.5% 3.0% 2.5% 2.0% 1.5% 1.0% 0.5% 0.0% Average Annual Rates ( ) Males Females Source: Human Mortality Database, HMD U.S. Population Mortality Improvements - Gender Average Annual Rates ( ) 6.0% 5.0% 4.0% 3.0% 2.0% 1.0% 0.0% -1.0% -2.0% -3.0% Males Females Source: Human Mortality Database, HMD 8

9 U.S. Population Mortality Improvements Geographic Region (Males) Average Annual Rates ( ) Northeast Midwest South West Grand Total 4.0% 3.5% 3.0% 2.5% 2.0% 1.5% 1.0% 0.5% 0.0% < 1 year Source: NCHS U.S. Population Mortality Improvements Geographic Region (Females) Average Annual Rates ( ) Northeast Midwest South West Grand Total 4.0% 3.5% 3.0% 2.5% 2.0% 1.5% 1.0% 0.5% 0.0% < 1 year Source: NCHS 9

10 U.S. Population Mortality Improvements Education Estimated Difference in Average Annual Mortality Improvement White Population, Ages Males Females 4.0% 3.0% 2.0% 1.0% 0.0% -1.0% -2.0% -3.0% -4.0% Less than 12 yrs 12 yrs yrs 16+ All Levels 19 Source: Widening Socioeconomic Inequalities in US Death Rates U.S. Population Mortality Improvements Education Estimated Difference in Average Annual Mortality Improvement African American Population, Ages Males Females 7.0% 6.0% 5.0% 4.0% 3.0% 2.0% 1.0% 0.0% -1.0% Less than 12 yrs 12 yrs yrs 16+ All Levels Source: Widening Socioeconomic Inequalities in US Death Rates 10

11 Cause of Death US General Population Rates Per 100,000 Population Pneumonia Tuberculosis Heart Disease Intracranial Lesions Nephritis Accidents Cancer Diabetes Source: CDC Reports Cause of Death Insured Population Limited Data Cancer declining Cardiovascular declining Accident declining Suicide increasing 22 11

12 Setting Mortality Improvement Assumptions for Life Insureds: Current Practices Insured Mortality Considerations Changes in lifestyle for target population exercise/diet Increase in obesity among adults and children in the US Smoking prevalence different impact than population Changes in underwriting paramedical testing, blood testing, uw rigor Changes in risk classifications smoker distinct, preferred classes Socioeconomic status of target population insureds tend to be in higher socioeconomic class than general population Impact of Lapses / Mortality Antiselection Insured data issues credibility of individual company data, industry studies - companies included not always consistent and varying product offerings, target markets, and distribution channels 24 12

13 U.S. Life Insureds Males Insured vs. Population Data Average Annual MI Rates ( ) General Population Life Insureds - Number of Pols Life Insureds - Amt 3.5% 2.5% 1.5% 0.5% -0.5% -1.5% 25 < 1 year Attained Age All Ages Source: SOA Studies of Standard Ordinary Life Insurance Mortality, HMD U.S. Life Insureds Females Insured vs. Population Data Average Annual MI Rates ( ) General Population Life Insureds - Number of Pols Life Insureds - Amt 3.5% 2.5% 1.5% 0.5% -0.5% < 1 year Attained Age All Ages 26 Source: SOA Studies of Standard Ordinary Life Insurance Mortality 13

14 Recent Insured Trends Ultimate Mortality Relatively low mortality improvement for ages < 50 Healthy improvements in mortality for ages 50+ Later duration experience is very good with ultimate mortality much improved at attained ages in the 50 s and 60 s 27 Recent Insured Trends Select Period Mortality More Difficult to Discern Mortality Trends Improvement for earliest durations Durations 3 and 4 who little or no improvement (antiselection?) 28 14

15 Growth in Prevalence of Obesity in the US Rates of Adult Obesity, Ages % 30% 31% 33% 35% 25% Pct of Population 20% 15% 10% 13% 15% 23% 5% 0% Source: CDC Data, NHANES (National Health and Nutrition Examination Surveys). Growth in Prevalence of Obesity in the US Rates of Childhood Obesity, Ages % 19% 20% 15% 16% Pct of Population 10% 11% 5% 7% 4% 0% Source: CDC Data, NHANES (National Health and Nutrition Examination Surveys). 15

16 Future Mortality Improvement Developing Insured Population Assumptions Data Sources General population data most common SSA data, HMD data, NCHS data General population data advantages Long experience period Consistency in underlying population data Detail available by socioeconomic and health factors from public sources Very limited credibility at the individual company level Does not have the noise caused by industry impacts, most significantly underwriting changes over time However, use of general population data introduces basis risk for insured assumptions. 31 Future Mortality Improvement Developing Insured Population Assumptions Methodology Considerations Historical MI vs future MI assumption Length of projection period Level of detail available in historical data (gender, age, health risk factors (smoking, obesity, drinking, exercise, etc.), cause of death, socioeconomic factors) Gender/Age Group currently the most common splits of future improvement assumptions 32 16

17 Future Mortality Improvement Developing Insured Population Assumptions Modeling Lee Carter and variants (Canadian study) CMI method and variants (blending to a long-term average) Cause-specific Disease-based Actuarial/Expert Judgment Study historical data Consider demographic drivers gender, age, socioeconomic factors Identify and track current and potential future drivers of change (ex. smoking prevalence, obesity, medical advances, etc.) 33 17