THE INSTITUTE OF ACTUARIES OF AUSTRALIA REPORT ON THE MORTALITY INVESTIGATION The IAA Mortality Committee

Transcription

1 THE INSTITUTE OF ACTUARIES OF AUSTRALIA REPORT ON THE MORTALITY INVESTIGATION By The IAA Mortality Committee C.M. Brennan BSc AIA FIAA P. Chan BSc FIAA A.P. Gale BSc FIAA A.T. Mak MEc FIAA P.R. Muir FIA FIAA ASA (Convenor) C.D. Murphy BSc(Hons) FIAA L.G. Tickle BEc FIAA G. Whittaker MA FIA FIAA J.Y. Wildgoose BSc (ex officio) Copyright Institute of Actuaries of Australia May 1999

2 Mortality Committee Report ACKNOWLEDGMENTS The Institute of Actuaries of Australia Mortality Committee wishes to thank the following organisations for their contributions of data to the IAA Mortality Investigation. Australian Casualty & Life Australian Eagle Life Australian Mutual Provident Society Commonwealth Life GIO Personal Investment Services MLC Life Mercantile Mutual Life Insurance Company The National Mutual Life Association of Australasia Norwich Union Life Australia The Prudential Assurance Company Royal & Sun Alliance Life Assurance Suncorp Insurance and Finance Tyndall Life Insurance Company The Institute also wishes to thank the following organisations for their financial support of the IAA Mortality Investigation. American International Assurance Company (Australia) Australian Casualty & Life Australian Eagle Life Australian Mutual Provident Society ANZ Life Assurance Company Citibank Cologne Life Reinsurance Company of Australia The Colonial Mutual Life Assurance Society Commonwealth Life FAI Life Insurance Society Gerling Global GIO Personal Investment Services Hallmark Life Insurance Company Hannover Life Re of Australasia Lumley Life MLC Life Mercantile Mutual Life Insurance Company Munich Reinsurance Company of Australia National Australia Bank The National Mutual Life Association of Australasia Norwich Union Life Australia NRMA Life The Prudential Assurance Company RGA Reinsurance Royal & Sun Alliance Life Assurance Suncorp Insurance and Finance Swiss Re Life & Health Tower Life Australia Tyndall Life Insurance Company Westpac Zurich Australian Life Insurance These financial and data contributions are in the public interest to support life insurance pricing and other practices. Finally, the Institute wishes to thank and acknowledge the fine effort by Jenny Wildgoose and also wishes to thank Macquarie University for the services provided since 1990, when the Investigation was redesigned on a non-profit commercial basis.

3 Mortality Committee Report 3 CONTENTS 1. EXECUTIVE SUMMARY 5 2. INTRODUCTION Basis for the report Previous reports and comparability Provision of data Disclaimer THE DATA Death Only Cover Trauma Cover Total and Permanent Disablement (TPD) Cover Immediate annuities Classification of the data for calculation of exposed-to-risk Data integrity RESULTS OF THE INVESTIGATION Policies with Death Only cover Trauma policies Total and Permanent Disablement (TPD) policies Immediate annuities BIBLIOGRAPHY 64 APPENDIX A: TABULATIONS OF THE DATA 65 A.1 Overview of all policy types for Death Only cover 66 A.2 Death Only cover: All non-annuity insurances with Death Only cover 68 combined A.3 Death Only cover: Whole of life and endowment insurances, with and 80 without term riders A.4 Death Only cover: Unbundled policies, both capital guaranteed and 86 investment-linked A.5 Death Only cover: Temporary insurances where premiums are fully 92 guaranteed

4 Mortality Committee Report A.6 Death Only cover: Temporary insurances where premiums may be 97 reviewed A.7 Female vs male experience: All non-annuity insurances combined, 103 Death Only cover A.8 Smoker-rated experience: All non-annuity insurances with Death Only 104 cover combined A.9 Smoker-rated experience: Term insurances with Death Only cover 112 A.10 Trauma cover: All non-annuity insurances combined 120 A.11 Female vs male experience: All non-annuity insurances combined, 126 Trauma cover A.12 Smoker-rated experience: All non-annuity insurances combined, 127 Trauma cover A.13 TPD cover: All non-annuity insurances combined 134 A.14 Female vs male experience: All non-annuity insurances combined, 140 TPD cover A.15 Smoker-rated experience: All non-annuity insurances combined, 141 TPD cover A.16 Immediate annuities 148 A.17 Female vs male experience: Immediate annuities 151 APPENDIX B: ESTIMATION OF STANDARD ERRORS IN ACTUAL TO 152 EXPECTED RATIOS APPENDIX C: SPECIFICATIONS TO THE MORTALITY 153 INVESTIGATION C.1 Introduction 153 C.2 Inclusions and Exclusions 155 C.3 Reporting of Data 156 C.4 Classification of Data 157 C.5 Supplied Edit Software 159 C.6 Changes to the Specifications 159

5 Mortality Committee Report 5 1. EXECUTIVE SUMMARY 1.1 Scope of the report This report, covering the calendar years 1995 to 1997, analyses the experience of Australian personal insurance business for both annuitant and non-annuitant mortality, and, for the first time, trauma benefits and total and permanent disability (TPD). Table 1.1: The volume of data included in the report by type of cover. Period of investigation is the three years TYPE OF COVER POLICIES YEARS CLAIMS/DEATHS EXPOSED-TO-RISK Death Only 4,500,000 20,496 Trauma 367, TPD 728, Annuity 25, Key features: Non-annuitant mortality Mortality improvements of about 1.5% p.a. for males and 1% p.a. for females from to , but with marked variation by policy type. Overall mortality for policies, durations 2 years and over, of 92% of IA M for males and 97% of IA F for females, again with marked variation by policy type. Evidence of a 5-year selection effect, especially for males. Female mortality at 62% of male mortality. A much higher smoker/non-smoker ratio for males (208%) than for females (138%). A wide variation in experience between offices, with the highest office at 1.5 to 3 times that of the lowest office, varying with policy type and duration. 1.3 Key features: Trauma The four major causes of trauma claims (cancer, coronary artery bypass surgery, heart attack and stroke) account for 84% of claims with cancer the most significant cause, especially for females. A selection effect is not apparent, particularly for females. Overall trauma experience from the four major causes, durations 2 years and over, of 44% and 48% of population morbidity for males and females respectively. Female trauma at about 87% of male trauma. Male smoker/non-smoker ratio is 166%; the equivalent ratio for females is not statistically credible.

6 Mortality Committee Report 1.4 Key features: TPD Strong evidence of a 5-year selection effect, especially for males. Female morbidity at about 61% of male morbidity; differing occupational mix by sex has not been analysed at this time, consequently occupational bias exists. Smoker/non-smoker ratios of 282% for males and 292% for females; use with caution due to very small numbers of smoker claims for both sexes. 1.5 Key features: Sums insured experience Actual to expected claims ratios by sums insured are lower than those by policy count for death, trauma and TPD, possibly reflecting the more stringent underwriting procedures for larger sums insured policies. 1.6 Key features: Annuities Mortality for annuitants, durations 1 year and over, is 58% of IM80 and 69% of IF80 for males and females respectively. Female mortality at about 75% of male mortality. Care is needed because of low exposed-to-risk and possible underreporting of deaths. Trends since the last three-year period are not clear due to low exposure and changes in the type of data collected. 1.7 Closing remarks The collection and analysis of insured lives experience is essential to the Australian life insurance industry and its clients. Areas of application of such analyses include pricing, reserving, underwriting, product design, solvency standards and international comparisons. In Australia the information provides the statistical justification for risk classification structures by, for example, age, sex and smoker status. The lower exposure in this report compared to the investigation is somewhat disappointing. A lower participation rate, compounded by the transition from the old to the new investigation format and the natural decline in the ageing conventional portfolio, has contributed to this. The Committee would like to urge the Appointed Actuaries of all life offices to continue (or begin) to contribute data in order to maintain the credibility and usefulness of this investigation. It is particularly important, firstly, to maintain consistent participation levels in order for meaningful time trends to emerge, and secondly, to achieve greater coverage of expanding markets, for example, annuities. Finally, the wide range of experience between offices and the low numbers of claims in some classes of business indicate that care is needed in using the aggregate industry results of this report.

7 Mortality Committee Report 7 2. INTRODUCTION 2.1 Basis for the report This report presents the final results of the Institute's investigation of the mortality of Australian personal (or individual ) lump-sum and annuity business in the three calendar years By using triennial publication we can provide an update to the previous investigation, indicating mortality and morbidity trends, despite the fluctuations that can occur from year to year in the claims from certain causes. The broad coverage of the investigation is of retail (as contrasted with wholesale) insured lives and annuitant business, and consequently most nonannuity policies have been underwritten. Whereas previous investigations have covered death only and annuity risks, this investigation comprehensively covers: trauma risks total and permanent disablement (TPD) risks the death element of policies with trauma risks (not covered in this report) the death element of policies with TPD risks (not covered in this report) policies which only cover the death risk ( death only ) immediate annuity policies (including those with or without a guaranteed period, but excluding term certain policies, and also including joint life annuities for the first time). The revised data specifications for this latest investigation period, which were changed to coincide with the commencement of the period, are attached as Appendix C. The Appendix includes descriptions of the various data codes and classifications used, for example, underwriting codes. Individual policy data have been collected for the first time. Previously, contributing offices provided summarised policy tabulations and individual claims. A more powerful computer has been used to edit and analyse the much greater volumes of data. Both the expansion of the types of risk collected, and also the introduction of individual policy data, have allowed more comprehensive analyses. Detailed analyses have been included as appendices, for convenience, and are commented upon and summarised in the body of the report. Wherever possible, requests from Institute members, at workshop seminars (particularly the November 1998 Life Insurance Forum), for new analyses have been included. However, more specific feedback is now sought from members so that the most useful analyses can be provided in future. The new investigation has been limited by the short period of time between the final close-off of data contributions and the publication date for the Darwin conference. Such industry phenomena as Year 2K and industry rationalisation have had their effect, as well as the effect on offices of the changes to specifications.

8 Mortality Committee Report We believe that the following achievements are worth highlighting: Publication of this report for standard lives is only 18 months after the close of the investigation period. Analyses are provided for the main product types and risks being sold in the Australian market by both policies and sums insured. Contributing companies are being provided with feedback analyses of their own office experience after each annual submission of data. The Committee has made plans for the following: An up-to-date insured lives graduated mortality table IA is planned for release later in 1999 to replace the previous table IA We believe that the cycle period for table replacement of five or six years is reasonable. In future the Institute will be able to decide whether to replace the published table on a three or six-year cycle. A separate analysis of non-standard (medically loaded) lives is planned to be released when sufficient data have been collected. 2.2 Previous reports and comparability The previous reports from the Mortality Committee which are relevant to this report are: REPORT RELEASE DATE Report on the Mortality Investigation March 1994 Mortality Investigation IA M and F, Release of Graduated Mortality Tables March 1994 Report on the Mortality Investigation June 1996 Sub-Standard Lives Experience Mortality Investigation May 1997 The full publication details of each report are provided in the Bibliography, 5. Analyses in this report have been prepared to provide results that can be compared with previous report analyses wherever possible, and thus indicate trends. However there are some limitations to complete comparability, and the following practical constraints should be noted. Product type popularity has waxed and waned (for example, whole life and endowment, unbundled, temporary insurances (premiums guaranteed), temporary insurances (premiums reviewable), trauma insurance). Because of this, selection effects differ for each product type, and comparisons between products require care. Contributing office data are not completely continuous as some offices have dropped out, others have commenced more recently and others have

9 Mortality Committee Report 9 omitted some years of data (please refer to previous reports and 2.3 of this report). Underwriting policy (and standards) change over time, and the mix of business sold and in-force is changing with regard to underwriting classifications (e.g. non-medical, medical, limited underwriting, and so on). Finally, the aggregate experience by policies (including all policy types and underwriting classifications) is not comparable with that for or However, subclassifications (for example, by product type) are reasonably comparable over time. The graduated IA M & F mortality table (IAA Mortality Committee, 1995b) has been used as a benchmark for measuring both the current insured lives mortality experience and the previous insured lives mortality experience. IM80/IF80 (Continuous Mortality Investigation, 1990) has been used for annuities. Comparability is facilitated by using these standard benchmarks. In addition, the scope of the investigation has been extended. Additional product types and underwriting categories have been included in the investigation, notably at 31 December 1994 when a range of new underwriting categories, including group continuation and guaranteed insurability options, conversions, and no underwriting, were added, and in November 1997 when master trust policies were added. The most recent period has analyses by sums insured. The annuitant investigation, which has been kept quite separate, includes a small group of joint life business in the latest period. When the IA95-97 graduated table is published, it is intended to provide additional comment and advice regarding its application to different product types and underwriting classifications. 2.3 Provision of data The report is based on data provided by thirteen offices. Six offices each had an average contribution exceeding 100,000 policies in force per census. Seven offices provided data for the full three year period, four offices for two of the three years and two for only one calendar year. Omissions were mainly due to teething problems in implementing the new data requirements.

10 Mortality Committee Report Table 2.3.1: Number of offices contributing data for each of the policy types in the report, by calendar year and in total for the three year period of investigation. NO. OF OFFICES BY YEAR NO. OF OFFICES OVER POLICY TYPE YEARS (i) Whole of life/endowment (ii) Unbundled (iii) Guaranteed term (iv) Reviewable term (v) Annuities It is hoped that in the future a greater market coverage can be achieved, as this would improve the significance of the results. The results have been based on standard lives, where "standard" refers to the assessment basis. Standard lives represent around 97% of all the in force business. 2.4 Disclaimer The Mortality Committee has endeavoured to ensure the accuracy of the data on which this report is based and to provide a useful interpretation of the results. Nevertheless, the Mortality Committee cannot accept any responsibility for errors or for the use or misuse of any information in this report. Actuaries and others using the results should exercise professional care and judgment in ensuring that they are suitable for the intended purpose.

11 Mortality Committee Report THE DATA In previous reports the statement has been made that the investigation is based on policies, rather than on lives. This statement, originally referring to the ownership of multiple policies by one life, is now a little ambiguous. Today, a significant number of policies have multiple lives attached to them, for example, joint annuities which normally have two lives and also a number of life policies where family groups are recorded under the one policy number. The Mortality Investigation has, in the case of life policies since 1989 and for annuities since 1994, allowed for the exposure of multiple lives under a single policy, provided that sufficient detail was held on each such life. Under the new specifications, offices submitted individual policy details, including policy type, sex, underwriting basis, sum insured, postcode, HIV testing status, smoking status, cover and loading for death, trauma and TPD risks, and occupation for TPD cover. Only one sum insured was collected for each policy; where more than one risk was covered, the sum insured collected was that for the first risk covered in order of death, trauma or TPD. This is a shortcoming, which may be corrected in future to provide an improved analysis by sums insured of trauma and TPD risks. For the purposes of this report, the data have been subdivided into the following risk categories: Policies with Death Only cover Policies with Trauma cover Policies with TPD cover Immediate annuities. While the first category excludes the second and third categories, the latter two categories are not mutually exclusive. This report covers standard lives only. A standard life is defined within each risk category as having zero loading for that risk. For example, a policy with Death Only cover is standard if it has zero medical loading for death. Of course, the concept of a standard life does not apply to annuity data. The policy types used in the investigation include the following: (i) Whole of life and endowment insurances, both with and without term insurance riders (ii) Unbundled policies, both of the capital guaranteed and investment-linked kinds, carrying significant death risk (iii) Temporary insurances where premiums are fully guaranteed; the sums insured may be level or reducing (iv) Temporary insurances where the premium rate may be reviewed (v) Immediate annuities.

12 Mortality Committee Report Individual Master Trust data were submitted by only one office and for reasons of confidentiality cannot be included in this report. When three or more offices are able to submit this type of data, an analysis of the aggregated Master Trust data will be produced. 3.1 Death Only Cover Death Only: The data by policy type and age Table shows the proportions of exposed-to-risk and claims over the period Whole of life and endowment policies, or policy type (i), constitute about 60% of policies exposed-to-risk, reduced from about two-thirds of the exposed-to-risk in the period, and almost 90% of claims. These policies continue to be an ageing portfolio with an average age, weighted by the exposed-to-risk, of 47 years, compared with average ages of 39 to 43 years for policy types (ii), (iii) and (iv). Moreover, 17% of these whole of life and endowment policies are for lives aged 60 or more, compared with less than 2% of unbundled and term policies. This preponderance of older lives in policy type (i) is clearly seen in Figure which shows the distribution of life policies by policy type and age. The other three life policy types are largely confined to the age range 15 to 65 years. The distribution for sums insured presents quite a different picture. Here, term insurances with reviewable premiums make up 66% of the exposed-to-risk and over 50% of the claims, reflecting the high level of sales in the last decade. Whole of life and endowment policies reduce to only 17% of the exposed-to-risk and 34% of the claims, largely an effect of the older lives, and relatively small sums insured, dominating this type of business which has been declining in popularity for many years. Table 3.1.1: Number and percentage of exposed-to-risk and claims for standard lives by policy type for Death Only cover. POLICY TYPE EXPOSED-TO-RISK % CLAIMS % Policies (i) 2,701, , (ii) 650, (iii) 87, (iv) 1,052, , Total 4,491, , Sums Insured $ $ (i) 48,998,312, ,079, (ii) 44,033,416, ,367, (iii) 4,755,051, ,581, (iv) 189,325,589, ,303, Total 287,112,369, ,332,

13 Mortality Committee Report 13 Figure 3.1.1: Numbers of policies exposed-to-risk, Death Only cover, by policy type and age in the period ,000 80,000 60,000 40,000 20, Age WOL/E REVIEW TERM UNBUNDLED GUAR TERM Figure 3.1.2: Sums insured exposed-to-risk ($millions), Death Only cover, by policy type in the period REVIEW TERM GUAR TERM UNBUNDLED WOL/E 0 100, ,000 Exposed-to-risk ($mills) The age distribution of the sums insured exposed-to-risk for policies with Death Only cover can be found in Appendix A, Tables A.2.7 to A.2.12.

14 Mortality Committee Report Death Only data by policy type and year of investigation Table shows the proportion of new policies by policy type for each year of the investigation. The relatively low statistics for 1995 are caused by the omission of data by some significant contributors. Table 3.1.2: POLICY TYPE Number and percentage of new policies exposed to risk (curtate duration 0) for standard lives, Death Only cover, by policy type and year of investigation EXPOSED-TO- % EXPOSED-TO- % EXPOSED-TO- RISK RISK RISK Policies (i) 1, , , (ii) 4, , , (iii) (iv) 48, , , Total 54, , , Sums insured (i) 100,647, ,736, ,378, (ii) 452,932, ,576, ,126,458, (iii) 55,564, ,114, ,167, (iv) 8,382,596, ,591,786, ,987,674, Total 8,991,741, ,803,213, ,609,678, %

15 Mortality Committee Report Death Only: The data by sex Data were subdivided by sex (Male, Female, and Undifferentiated). Policies classified as Undifferentiated are almost entirely older policies on male lives. For the purpose of this report the Undifferentiated category has been included with Males. Table 3.1.3: Number and percentage of exposed-to-risk, Death Only cover, and death claims for standard lives by sex and policy type POLICY TYPE EXPOSED-TO-RISK % CLAIMS % Male Policies (i) 2,024, , (ii) 450, (iii) 48, (iv) 647, Total 3,171, , Female Policies (i) 677, , (ii) 200, (iii) 39, (iv) 404, Total 1,320, , Male Sums insured (i) 36,481,970, ,284, (ii) 31,989,545, ,128, (iii) 2,895,898, ,602, (iv) 130,432,873, ,120, Total 201,800,287, ,136, Female Sums insured (i) 12,516,341, ,794, (ii) 12,043,870, ,239, (iii) 1,859,153, ,979, (iv) 58,892,716, ,182, Total 85,312,082, ,195,

16 Mortality Committee Report Death Only: The data by underwriting basis Data were also subdivided by underwriting basis (Minimum Evidence, Non- Medical, Medical, Para-Medical, Group Continuation Option, Guaranteed Insurability Option, Conversion, None, and Unknown). While most offices provided the underwriting basis for most policies, two offices were not able to provide the underwriting basis at all. This accounts for the large proportion of policies coded as Unknown. If we can assume that these policies with unknown underwriting basis have much the same underwriting distribution as policies where the underwriting basis has been coded, the proportion of policies with either Non-Medical or Medical underwriting rises to over 95% by both numbers of policies and sums insured. Table 3.1.4: Percentage of exposed-to-risk, Death Only cover, and claims for standard lives by underwriting basis, policy type, and sex % EXPOSED-TO-RISK UNDERWRITING POLICIES SUMS INSURED BASIS MALES FEMALES MALES FEMALES Min Evidence Non-Medical Medical Para-Medical Group Cont Guar Insurability Conversion None Unknown TOTAL

17 Mortality Committee Report Death Only: The data by HIV testing status Data were subdivided by HIV testing status (Tested, Not Tested, and Unrecorded). HIV testing has been gradually introduced for non-annuitant products from about Table shows the proportions of recent policies by the HIV testing status. It is clear that a number of offices are finding this information too hard to code adequately. The high proportion of reviewable term policies (policy type (iv)) in the Unrecorded category is particularly surprising, given the relatively large sums insured in this policy type. No use has been made of HIV testing information in the results. Table 3.1.5: Percentage of standard policies in force with an entry date in 1995 or later by HIV testing status, policy type, and sex POLICY TYPE TESTED NOT TESTED UNRECORDED TOTAL % % % % Male policies (i) (ii) (iii) (iv) TOTAL Female policies (i) (ii) (iii) (iv) TOTAL Male sums insured (i) (ii) (iii) (iv) TOTAL Female sums insured (i) (ii) (iii) (iv) TOTAL

18 Mortality Committee Report Death Only: The data by smoking status Data were subdivided by smoking status (Smoker, Non-Smoker, and Aggregate Rating). Both the Table (by policy count) and Table (by sums insured) provide clear evidence of the trend to smoker rating, particularly for term insurances (types (iii) and (iv)), but also for unbundled business (type (ii)). Traditional business and its term riders, however, are less commonly smoker rated. Table 3.1.6: POLICY TYPE Numbers of policies exposed-to-risk, Death Only cover, and percentage by smoking status, policy type and sex, for both recent and older standard policies EXPOSED- TO-RISK SMOKER NON- AGGREGATE TOTAL SMOKER % % % % Males Durations 0 to 4 (i) 77, (ii) 98, (iii) 9, (iv) 467, TOTAL 653, Males Durations 5 and over (i) 1,946, (ii) 351, (iii) 38, (iv) 180, TOTAL 2,517, Females Durations 0 to 4 (i) 57, (ii) 51, (iii) 6, (iv) 285, TOTAL 401, Females Durations 5 and over (i) 620, (ii) 148, (iii) 32, (iv) 118, TOTAL 919,

19 Mortality Committee Report 19 Table 3.1.7: POLICY TYPE Sums insured exposed-to-risk ($), Death Only cover, and percentage by smoking status, policy type and sex, for both recent and older standard policies EXPOSED-TO- SMOKER NON- AGGREGATE TOTAL RISK SMOKER $ % % % % Males Durations 0 To 4 (i) 5,401,714, (ii) 9,412,338, (iii) 757,716, (iv) 100,493,640, Total 116,065,410, Males Durations 5 and over (i) 31,080,255, (ii) 22,577,206, (iii) 2,138,181, (iv) 29,939,232, Total 85,734,876, Females Durations 0 to 4 (i) 2,715,696, (ii) 4,237,960, (iii) 480,725, (iv) 44,809,503, Total 52,243,884, Females Durations 5 and over (i) 9,800,645, (ii) 7,805,910, (iii) 1,378,427, (iv) 14,083,213, Total 33,068,197,

20 Mortality Committee Report Death Only: The data by assessment basis The analyses in this report are based on standard lives only. Medically substandard lives will be analysed in a subsequent report, necessarily over a longer period due to the scantiness of the data. It is interesting to note, in Table 3.1.8, the lower proportion of standard business, and thus the higher proportion of medically loaded business, by sums insured. This is not unexpected due to the higher underwriting standards applied by direct companies and their reinsurers to larger sums insured (medical business). Table 3.1.8: Percentage of exposed-to-risk for all lives by assessment basis and sex % EXPOSED-TO-RISK ASSESSMENT POLICIES SUMS INSURED BASIS MALES FEMALES MALES FEMALES Standard Light Medium Heavy TOTAL Key: Standard Zero loading Light 0 < Loading 50% Medium 50 < Loading 100% Heavy 100% < Loading

21 Mortality Committee Report Death Only: The data by duration The risk commencement date was supplied for each policy and claim. Table clearly shows the changing mix of sales by policy type. While policy types (i), (ii) and (iii) have 75% or more of policies exposed-to-risk at duration 5 and over, less than 30% of policy type (iv) are at these longer durations. Table 3.1.9: Percentage of exposed-to-risk, Death Only cover, for standard lives by policy type, sex and duration POLICY PERCENTAGE BY DURATION TOTAL TYPE % Male Policies (i) (ii) (iii) (iv) TOTAL Female Policies (i) (ii) (iii) (iv) TOTAL Male Sums Insured (i) (ii) (iii) (iv) TOTAL Female Sums Insured (i) (ii) (iii) (iv) TOTAL

22 Mortality Committee Report 3.2 Trauma Cover Trauma: The data by policy type and age Table shows the exposed-to-risk and claims by policy type for both policy count and sums insured. The predominance of policy type (iv) is remarkable. The analyses of trauma cover by sums insured should be treated with caution due to the death sum insured being used as a proxy for the actual trauma sum insured in those policies carrying both death and trauma risks, a large proportion of the data. Table 3.2.1: Number and percentage of exposed-to-risk, Trauma cover, and trauma claims for standard lives by basic policy type. POLICY TYPE EXPOSED-TO-RISK % CLAIMS % Policies (i) 4, (ii) 4, (iii) (iv) 358, TOTAL 367, Sums Insured $ $ (i) 446,033, , (ii) 470,852, , (iii) 767, (iv) 65,034,742, ,589, TOTAL 65,952,395, ,987, Not surprisingly, the exposed-to-risk is concentrated in the age range 17 to 65 years, because these are the issue restrictions and expiry ages imposed by writers of trauma business in Australia. There is a very small percentage exposed at ages 0 to 16, less than 0.5%; these are special trauma policies, insuring children, now offered by some companies. The majority of lives are aged years. Claims occurring at ages years account for about 80% of the total number for all events combined. Table shows the proportions of exposed-to-risk and total claims for all events by sex in each age band for

23 Mortality Committee Report 23 Table 3.2.2: Percentage of distribution of total exposed-to-risk and claims (all events), Trauma cover, by age and sex, POLICY TYPE EXPOSED-TO-RISK % CLAIMS % Males , , , , , TOTAL 205, Females , , , , , TOTAL 161, Persons , , , , , TOTAL 367,

24 Mortality Committee Report Figure 3.2.1: Total exposed-to-risk for all non-annuity policies, Trauma cover, by age in the period ,000 Exposed-to-risk 20,000 10, Age Trauma: The data by year of investigation Table shows the exposed-to-risk and number of claims for each year of the investigation period that are included in the analysis. This report covers the collective trauma experience of nine companies. Three companies (with a total of 117 claims) have been omitted from the investigation, because they did not record specific cause of claim. Four companies were able to supply data for all three years of the investigation, another four companies for two out of the three years and one company for 1997 only. Table 3.2.3: Number of policies exposed to risk and claims by year of investigation EXPOSED- CLAIMS TOTAL YEAR TO-RISK 4 MAJOR EVENTS* OTHER EVENTS CLAIMS , , , TOTAL 367, * Includes Heart Attack, Cancer, Stroke and Bypass Surgery.

25 Mortality Committee Report Trauma: The data by sex Data were subdivided by sex (Male, Female, and Undifferentiated). The very small number of policies classified as Undifferentiated were grouped with Males. Of the policies investigated, 56% are male and 44% are female. Claims are similarly distributed in number. The male proportion of exposed-torisk is higher (65%) at durations 10 and over. Table 3.2.4: Numbers of exposed-to-risk, Trauma cover, and all trauma claims for standard lives by sex EXPOSED-TO-RISK % CLAIMS % Policies Male 205, Female 161, TOTAL 367, Sums insured Male 41,522,182, ,837, Female 24,430,213, ,150, TOTAL 65,952,395, ,987,

26 Mortality Committee Report Trauma: The data by smoking status Data were subdivided by smoking status (Smoker, Non-Smoker, and Aggregate Rating). Table shows both the heavy recent sales and the trend to almost all business being smoker rated. About 85% of policies are non-smokers with a similar proportion observed for the total number of claims. Table 3.2.5: Numbers of policies and sums insured exposed-to-risk, Trauma cover, and percentage by smoking status and sex, for both recent and older standard policies SEX AND EXPOSED-TO- SMOKER NON- AGGREGATE TOTAL CURTATE RISK SMOKER DURATION % % % % Policies: Males Duration , Duration 5+ 6, Policies: Females Duration , Duration 5+ 5, Sums insured: Males Duration ,271,939, Duration 5+ 1,250,243, Sums insured: Females Duration ,745,659, Duration ,553,

27 Mortality Committee Report Trauma: The data by duration The risk commencement date was supplied for each policy and claim. The data were subdivided by the curtate duration at the census date for policies exposed-to-risk and for claims at the date of the trauma event or the date of diagnosis where there is no clearly defined event. Trauma business first appeared in Australia from the mid-1980s and is showing tremendous growth in new business, clearly seen in Table below. Table 3.2.6: Percentage of exposed-to-risk, Trauma cover, for standard lives by sex and duration POLICIES SUMS INSURED DURATION MALE % FEMALE % MALE % FEMALE % < 3 months to 6 months to 12 months to 2 years to 3 years to 4 years to 5 years to 10 years years TOTAL Trauma: The data by cause of claim The distribution of trauma claims by major event and sex is shown in Table Table 3.2.7: Number of trauma claims by sex and major event ACTUAL NUMBER OF CLAIMS BY SEX AND EVENT MAJOR EVENT MALES FEMALES PERSONS Cancer Heart Attack Bypass Surgery Stroke TOTAL

28 Mortality Committee Report 3.3 Total and Permanent Disablement (TPD) Cover TPD: The data by policy type and age Table shows the exposed-to-risk with TPD cover and TPD claims for both policy count and sums insured by policy type. As with trauma cover, the analyses of TPD cover by sums insured should be treated with caution due to the death sum insured being used as a proxy for the actual TPD sum insured in those policies carrying both death and TPD risks. Figure shows that the distribution of TPD policies by age is very similar to that for trauma cover. Table 3.3.1: Number and percentage of exposed-to-risk, TPD cover, and TPD claims for standard lives by policy type. POLICY TYPE EXPOSED-TO-RISK % CLAIMS % Policies (i) 115, (ii) 222, (iii) 26, (iv) 364, TOTAL 728, Sums insured (i) 3,579,070, ,404, (ii) 15,856,133, ,231, (iii) 2,108,021, ,360, (iv) 74,387,118, ,444, TOTAL 95,930,344, ,440, Figure 3.3.1: Total exposed-to-risk for standard lives, TPD cover, by age in the period ,000 30,000 Exposed-to-risk 20,000 10, Age

29 Mortality Committee Report TPD: The data by sex Table 3.3.2: Numbers of exposed-to-risk, TPD cover, and all TPD claims for standard lives by sex. EXPOSED-TO-RISK % CLAIMS % Policies Male 497, Female 230, TOTAL 728, Sums insured Male 70,377,387, ,564, Female 25,552,956, ,875, TOTAL 95,930,344, ,440, TPD: The data by smoking status Data were subdivided by smoking status (Smoker, Non-Smoker, and Aggregate Rating). Most new TPD sales are now smoker rated. However, there is more longer duration business for TPD than for trauma cover, as TPD has been sold for many years. Table 3.3.3: SEX AND DURATION Percentage of exposed-to-risk, TPD cover, by smoking status and sex, for both recent and older standard policies. EXPOSED-TO- RISK SMOKER NON- AGGREGATE TOTAL SMOKER % % % % Male Policies Duration , Duration , Female Policies Duration , Duration 5+ 90, Male Sums insured Duration ,713,460, Duration 5+ 18,663,927, Female Sums insured Duration ,590,606, Duration 5+ 5,962,349,

30 Mortality Committee Report TPD: The data by duration The risk commencement date was supplied for each policy and claim. Table shows that TPD cover remains a reasonably popular product. Table 3.3.4: Percentage of exposed-to-risk, TPD cover, for standard lives by sex and duration. PERCENTAGE BY DURATION TOTAL % Policies Male Female Sums Insured Male Female

31 Mortality Committee Report Immediate annuities Annuitant data by age It can be seen from Figure that the number of annuitants begins to rise steeply from about age 60, peaks around age 70, and thereafter falls rapidly, flattening off after age 80. This very prominent peak at age 70 is due not to a sudden increase in mortality at this age, but to the increased popularity of annuity products amongst recent retirees. Table 3.4.1: Number of annuitants exposed-to-risk and deaths by count and annual amount of annuity ($). POLICY TYPE EXPOSED-TO-RISK DEATHS Annuitants 24, Amount of annuity $ 185,862,446 4,128,341 Figure 3.4.1: Annuitants exposed-to-risk by age. 2,500 2,000 Exposed-to-risk 1,500 1, Age

32 Mortality Committee Report Annuitant data by annuity type and sex The data were subdivided by the annuity type (Single Life, Joint Life Last Survivor and Joint Life Reversionary). In Table 3.4.2, the predominance of Joint Life business is clear for the first time, even after one office s Joint Life annuities were excluded due to data problems. The number of female annuitants is now larger than the number of males. Table 3.4.2: Annuitants and annual amount of annuity ($) exposed-to-risk by annuity type and sex. ANNUITY POLICIES AMOUNT OF ANNUITY $ TYPE MALE % FEMALE % MALE % FEMALE % Single Life Joint Life Last Survivor Joint Life Reversion 4, , ,921, ,361, , , ,424, ,595, ,237, ,322,743 4 TOTAL 11, , ,583, ,279,

33 Mortality Committee Report Classification of the data for calculation of exposed-to-risk Offices supplied individual claims notified each quarter with date of birth, risk commencement date and date of death. The In Force was supplied in the form of a census of individual policies in force at the close of business on 31 December. One office supplied in force data as at 30 June and this was taken into account when calculating the exposed-to-risk. Care has been taken in allowing for late notifications of claims. Claims have been recorded by age attained at the date of death and where a census intervened between the date of death and the date of notification the in force has been adjusted appropriately. 3.6 Data integrity It is of utmost importance to ensure that the data are credible and as free from error as possible. Numerous checks have been applied to the data to this end. a) A validation program was issued to all contributing offices to allow them to remove any gross errors before submitting the data to the Mortality Investigation. The program produces an error listing and also a set of tables summarising the data by each major categorisation. b) The supervising actuary for each office was required to certify the summary tables produced by the validation program as being a reasonable representation of the office's data. c) A policy identifier was included in each individual policy record to allow duplicate records to be detected and referred back to the contributing office for checks on, for example, possible misclassification. d) Policies appearing for the first time at a census which, according to the risk commencement date, should have appeared at an earlier census are referred back to the contributing office. This has proved to be a valuable check on understatement of policies at duration zero. It has also been useful in identifying blocks of business which appear in a census for the first time, without having been included in a previous census. e) As the Mortality Investigation is based on first claims only, policies which are reported as in force after a claim has occurred are removed from the in force. f) Claims which, according to the risk commencement date, should have an earlier in force record, but do not, are checked. If necessary, the missing in force records are reinstated at the appropriate census(es). g) Claims are also checked for possible misclassification against the previous in force. h) In each census of annuities the classification of joint annuitant records is checked. Any inconsistencies are referred back to the contributing office.

34 Mortality Committee Report i) The computer system devised for the Mortality Investigation tracks the history of each office's data, recording the details of each submission, including the policy types represented at each census. This history is used by the computer system to control the calculation of the exposed-to-risk and total claims for a given period of investigation. The period is divided into its component calendar years and the exposed-to-risk and claims are calculated for each year individually. An office is included in the calculation for the calendar year, if a census has been submitted for both the census date beginning the year and the census date ending the year and claims have been submitted for all quarters in between. For each office included in a calendar year, policy types not represented at both census dates are excluded from the calculation of both the exposed-to-risk and the total claims for that year. j) A strategy was developed to manage the problems posed for the Mortality Investigation by company mergers. This same strategy was also applied to situations where a company wished to add a class of business to their submission which had not been included in past submissions. The company concerned was requested to submit two sets of in force data for the census date at which the changeover would take effect, one in force to end the old year and the other to start the new. k) Standard Individual Office Reports in spreadsheet form have been compiled for each contributing office on receipt of each year s data. These have proved to be valuable feedback to the contributors and also provide a further opportunity for offices to check their data. As part of the normal data integrity checking and production of the Standard Individual Office reports, errors in some offices data extraction programs were detected and, in most cases, were able to be rectified in time for the erroneous data to be re-submitted. Two offices were unable to fix problems with their annuity data in time for this report; one office s annuities were completely excluded due to a large number of unreported deaths, the other office s joint annuity data was excluded due to many missing in force records. Overall, the change to the submission of in force data individually rather than in summary form and the provision of analyses back to contributing offices have produced an improvement in the data integrity of the investigation. Further improvements are expected in the future.

35 Mortality Committee Report RESULTS OF THE INVESTIGATION 4.1 Policies with Death Only cover Comparison of volume of data with earlier investigations Tables A.1.1 and A.1.2, in Appendix A show the volume of exposed-to-risk and claims for all standard policies in this investigation. The following table, Table 4.1.1, compares the volume of non-annuity data, Death Only cover, for standard male lives at durations 2 and over in the Institute s , and investigations. Policies with Trauma or TPD cover were not included in the Mortality Investigation prior to The volume of male non-annuity exposure data in the investigation for is 54% of that used in the investigation, and 57% of that used in the investigation. The male claim numbers in the investigation for are also correspondingly lower than in the previous two investigations: 54% of and 58% of All three investigations cover a three year period. The number of contributing offices has risen by one, but this has been more than offset by two factors. Firstly, one significant life office was not able to contribute data. Secondly, five contributing companies were not able to supply 1995 data, due mainly to the change in data specifications, effective from Table 4.1.1: POLICY TYPE Numbers and percentages of life policies exposed-to-risk and claims, Death Only cover, for standard male lives at durations 2 and over in the , and periods of investigation. EXPOSED- TO-RISK % CLAIMS % (i) 3,943, , (ii) 720, , (iii) 204, (iv) 174, Total (i)-(iv) 5,042, , (i) 3,865, , (ii) 1,048, , (iii) 118, (iv) 288, Total (i)-(iv) 5,321, , (i) 2,003, , (ii) 421, (iii) 46, (iv) 408, Total (i)-(iv) 2,878, ,