Professeur, École d’actuariat

Slides:



Advertisements
Similar presentations
Annual Meeting ● Assemblée annuelle 2007 Vancouver 2007 Annual Meeting ● Assemblée annuelle 2007 Vancouver Canadian Institute of Actuaries Canadian.
Advertisements

1 1 Slide IS 310 – Business Statistics IS 310 Business Statistics CSU Long Beach.
Inference for Regression
1 1 Slide STATISTICS FOR BUSINESS AND ECONOMICS Seventh Edition AndersonSweeneyWilliams Slides Prepared by John Loucks © 1999 ITP/South-Western College.
1 Winter 2014 Building Blocks – York’s Pension Plan.
The Changing Well-being of Older Status First Nations Adults An Application of the Registered Indian Human Development Index Symposium on Aboriginal Experiences.
© 2010 Pearson Prentice Hall. All rights reserved Least Squares Regression Models.
A neural network based several-hour- ahead electric load forecasting using similar days approach Paras Mandal, Tomonobu Senjyu, Naomitsu Urasaki, Toshihisa.
Chapter 3 Forecasting McGraw-Hill/Irwin
Chapter Topics Types of Regression Models
Forecasting McGraw-Hill/Irwin Copyright © 2012 by The McGraw-Hill Companies, Inc. All rights reserved.
Short Term Load Forecasting with Expert Fuzzy-Logic System
Compare Outcomes Using all the above specific categories, we could compare 0-4 year-old male Asian mortality rates for asthma with 0-4 Asian female rates.
1 POPULATION PROJECTIONS Session 6 - Introduction to population projections Ben Jarabi Population Studies & Research Institute University of Nairobi.
Business Forecasting Chapter 4 Data Collection and Analysis in Forecasting.
DDSS2006 A Comparison Study about the Allocation Problem of Undesirable Facilities Based on Residential Awareness – A Case Study on Waste Disposal Facility.
T-TEST Statistics The t test is used to compare to groups to answer the differential research questions. Its values determines the difference by comparing.
Chapter 17 Time Series Analysis and Forecasting ©.
1 Modeling Coherent Mortality Forecasts using the Framework of Lee-Carter Model Presenter: Jack C. Yue /National Chengchi University, Taiwan Co-author:
1 Mortality Compression and Longevity Risk Jack C. Yue National Chengchi Univ. Sept. 26, 2009.
Lecture 8 Simple Linear Regression (cont.). Section Objectives: Statistical model for linear regression Data for simple linear regression Estimation.
MC Chapter 27 – Inference on slope. Growth hormones are often used to increase the weight gain of chickens. In an experiment using 15 chickens, five different.
Accuracy Based Generation of Thermodynamic Properties for Light Water in RELAP5-3D 2010 IRUG Meeting Cliff Davis.
Helmholtzstraße 22 D Ulm phone+49 (0) 731/ fax +49 (0) 731/ It Takes Two: Why Mortality Trend Modeling is more.
Transition costs and their impact on adequacy Vidija Pastukiene Seminar on Private Pension Provision Transition costs and decumulation phase Tallinn, 6-7.
A Stochastic Model of CPP Liabilities – Preliminary Results Rick Egelton Chief Economist CPPIB October 27, 2007 The views in this presentation reflect.
© 2008 Morningstar, Inc. All rights reserved. 3/1/2008 LCN Role of Immediate Annuities in Retirement.
“The Effect of Changing Exposure Levels on Calendar Year Loss Trends” by Chris Styrsky, FCAS, MAAA Ratemaking Seminar March 10, 2005.
Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.
Dynamic Models, Autocorrelation and Forecasting ECON 6002 Econometrics Memorial University of Newfoundland Adapted from Vera Tabakova’s notes.
Chapter 14 Introduction to Regression Analysis. Objectives Regression Analysis Uses of Regression Analysis Method of Least Squares Difference between.
Prepared for the Annuity Reserve Work Group By Steve Strommen FSA, CERA, MAAA May 1, 2013 POTENTIAL RESERVE METHODOLOGY.
Design and Data Analysis in Psychology I Salvador Chacón Moscoso Susana Sanduvete Chaves School of Psychology Dpt. Experimental Psychology 1.
“The Effect of Changing Exposure Levels on Calendar Year Loss Trends” by Chris Styrsky, FCAS, MAAA MAF Seminar March 22, 2005.
Fundamentals of Data Analysis Lecture 10 Correlation and regression.
Chapter 13 Simple Linear Regression
Inferential Statistics
Statistical Significance
2012-based Population Projections for England
The Burden of Colorectal Cancer in Arkansas
U.S. Social Security Administration Mortality Projections
LIABILITY PORTFOLIO MANAGEMENT Diversification of Longevity and Mortality Risk Stuart Silverman, FSA, MAAA, CERA Longevity 12 September 29-30, 2016.
Steve Goss, Chief Actuary U.S. Social Security Administration
Statistical Quality Control, 7th Edition by Douglas C. Montgomery.
business analytics II ▌assignment one - solutions autoparts 
SOA Annual Meeting 2005 Fertility Symposium
POPULATION PROJECTIONS
SOA Annual Meeting 2005 Fertility Symposium
Linear Regression and Correlation Analysis
APPROACHES TO QUANTITATIVE DATA ANALYSIS
Estimating with PROBE II
Project Sales or Production Levels Using the Rolling Average
ARITHMETIC INCREASE METHOD
CIA meeting Group Hot Topics SOA 2005 Waiver Table
Insurance IFRS Seminar December 1, 2016 Darryl Wagner Session 17
Insurance IFRS Seminar December 1, 2016 Darryl Wagner Session 17
Sample Size and Accuracy
Comparing Populations
Statistics Time Series
Logistic Regression.
Labour Price Index Labour Market Statistics (LAMAS) Working Group
Click the mouse button or press the Space Bar to display the answers.
SIMPLE LINEAR REGRESSION
Design and Data Analysis in Psychology I
Figures adapted from the TIEDI Analytical Report #4:
SIMPLE LINEAR REGRESSION
Diagnostics and Remedial Measures
City Council Meeting May 23, 2011
Table 2. Regression statistics for independent and dependent variables
Diagnostics and Remedial Measures
Presentation transcript:

Professeur, École d’actuariat Louis Adam 2006-09-22 Session # PD-11 CLIFR Topics II Expected Annuitant Mortality Improvement Louis Adam, f.s.a., f.i.c.a. Professeur, École d’actuariat Université Laval Toronto, September 22, 2006 Temps alloué de 1h30 De 10h30 à 12h00 I will focus on results of a reseach project involving only Canadian data for pensioners. CIA AA Seminar, PD 11 CLIFR Topics II Mortality Improvement

Outline Introduction: Modified AA scale OK ? Trend of Pensioner Mortality in Canada Proposed Projection Scale 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

Part 1- Introduction Conclusion to be announced Modified Scale AA Proposed scale 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

Announced conclusion Proposed scale is higher than Modified AA scale (see next slides for conclusion) Results will show how figures are achieved Methodology debatable, not the data No crystal ball for the future Mirror-image: Short-term future assumed to mimic recent past 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

Modified Scale AA (CLIFR, October 2005) 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

Projection Scale: Male 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

Projection Scale: Female 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

Reference for Scale AA « Scale AA is wholly based on the historic experience of the CSRS and of Social Security--1977 through 1993, by age and sex--with a minimum 0.5% per year improvement at ages under 85. The period 1977 through 1993 was considered a representative historical period on which to base future mortality projections. » THE UP-94 AND GAR-94 TABLES: ISSUES IN CHOOSING THE APPROPRIATE TABLE, TSA XLVII, 1995, page 797 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

Issues with Scale AA U.S. data only Experience ends in 1993 Current standard Is it similar to recent Canadian experience? 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

Modified AA Scale CIA, CLIFR, October 2005 Changes at various ages, from AA scale Female: minimum rate of 1 % from age 60 to age 80 Still below recent Canadian experience, according to this data set 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

Proposal for a new scale Canadian data 1967-2003 Period CPP + QPP Pensioners (over age 60) Only those with pensions > 35 % Maximum Pension ( exclude low earnings) Average rate of decrease, with adjustment 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

New scale features Proposed figures based on past Canadian experience Future experience might vary Slight differences in point estimates based on specific method of calculation Confidence intervals shown for one method 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

Part 2- Mortality Trend Projection Scale: q70 ×(1-Rate70)#years Weights given to each year Length of period ending 2003 10 years (5-10-15-20-25 ??) Maximum allowed by data ? Implicit assumption: future trend similar to “a part of the past experience” 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

Evolution of q70 over time 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

Detailed calculation Trend over past 10 years Annual decrease of q70 Probabilities, rate of decrease, weights Rate of decrease is not constant over time: recent acceleration Arithmetic average of weighted rate of decrease 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

2003 10-year Rate, q70 Explain how 2.98 % is obtained q1994 / q1993 = 1 - rate (1994) Weight(1994) = 1 / Var(q1994) Calculate rate(1994)… rate(2003) 2.98 % = Weighted average 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

2003 10-yr AA 70 for q70 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

2003 10-yr Least-square Rate 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

Weighted Rate of Decrease 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

Long-Term Rate Average over longest possible period Vary by age (C/QPP inception, etc.) Age 60: from 1984 to 2003, 20 values. 70: 1967 – 2003, 37 values 80: 1976 – 2003, 28 values 90: 1986 – 2003, 18 values 100: 1996 – 2003, 8 values 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

Alternate values for AA70 Louis Adam 2006-09-22 Alternate values for AA70 Weighted avg.: 2.98 % Unweighted average of 10 figures from 1994 to 2003 : 2.76 % Least-square, unweighted: 2.97 % Least-square, weighted: 3.02 % 95 % Confidence Interval: From 2.49 % to 3.55 % (± 0.5 % ) R2 = 94.7 % This is important ! : significant difference with AA scale. 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement CIA AA Seminar, PD 11 CLIFR Topics II Mortality Improvement

Alternative methods Unweighted least squares Example: 15-yr rate, q70 Arithmetic avg, unweighted: 2.53 % Arithmetic avg, weighted: 2.78 % Least Squares (next slide): 2.46 % Least Square, weighted: 2.54 % 95% Confidence Interval: 2.21 % to 2.88 % (± 0.3 % ) Mortality decreases over time 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

Unweighted Rate: 2.46 %, Weighted Rate: 2.54 % (15-year) 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

Impact of Confidence Interval Next slides show, for various time-periods: Weighted least-square regression avg rate of decrease of mortality Lower and upper bounds for rate of decrease Modified Scale AA My 2005 Proposal (based on arithmetic avg) 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

95 % Confidence Interval: 10-year, Male 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

95 % Confidence Interval: Maximum Period, Male 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

95 % Confidence Interval: 25-Year Period, Male 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

95 % Confidence Interval: 10-year, Female 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

95 % Confidence Interval: Maximum Period, Female 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

95 % Confidence Interval: 25-Year Period, Female 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

Male, Various Periods: Mean 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

Female, Various Periods: Mean 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

Male Projection Factors 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

Female Projection Factors 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

Projection Scale: Male 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

Projection Scale: Female 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement

Some comments Does qx decrease over time ? Yes ! Louis Adam 2006-09-22 Some comments Does qx decrease over time ? Yes ! Rate of decrease not constant by age Higher at age 65, lower at age 90 Negative decrease over age 90 ? Varies from year to year Rates for female lower than male Recent rates of decrease are higher than in the past Proposal is relatively simple (straight line sought) Will the future be like the past ? Is it applicable to your block of business ? 2006-09-22 Louis Adam, CIA AA Seminar, PD 11 CLIFR Topics II, Mortality Improvement CIA AA Seminar, PD 11 CLIFR Topics II Mortality Improvement

Thank you ! Questions ?