Study. Study on Canadian Group Long Term Disability Termination Experience ( ) Group Life and Health Subcommittee Research Committee

Transcription

1 Study Study on Canadian Group Long Term Disability Termination Experience ( ) Group Life and Health Subcommittee Research Committee January 2010 Document Ce document est disponible en français 2010 Canadian Institute of Actuaries

2 Table of Contents EXECUTIVE SUMMARY 3 ACKNOWLEDGEMENTS 5 INTRODUCTION 6 GROUP DISABILITY CHARACTERISTICS AND PLAN DESIGN 7 CONFIDENTIALITY 8 DATA COLLECTING AND REFINEMENT METHODOLOGY 8 DATA SUMMARY 10 GRADUATION OF DATA 16 APPENDIX A DATA SUMMARIES 18 APPENDIX B COMPARISON TO PUBLISHED TABLES 25 APPENDIX C ILLUSTRATION OF THE GRADUATED RESULTS 33 APPENDIX D GRADUATED TABLES 40 APPENDIX E GRADUATION FORMULAS 48 2

3 EXECUTIVE SUMMARY This is an update of the earlier termination study done by the Research Committee - Group Life & Health Subcommittee (the Committee). This study includes data from some additional insurers as well as data for the 1996 and 1997 years. Also included in this report are a number of graduated tables. The various tables are: Disabled recovery (Québec/Non-Québec, unisex) Disabled mortality (Québec/Non-Québec, gender specific) The graduated tables that have been produced reflect the average experience for the period and do not include any margins. While the information coming out of this study gives an indication of recent aggregate Canadian experience, the Committee did not analyze the data by other factors such as industry, occupation, benefit level, plan design or tax status. We do understand that these factors may contribute to the experience we believe this detail is best left for private studies. The Committee did attempt to separate the data based on whether or not the individual had a CPP/QPP award or not. However, this field was only populated in less than 40% of the records. Data were collected from virtually all insurers in Canada, covering the period 1988 to The data collected represent over 300,000 records and approximately 95% of the LTD premium in Canada. The data were provided by the 25 companies listed below, which together receive an estimated 95% of the LTD premiums in Canada. Many of these companies have subsequently merged. Aeterna Vie Aetna Canada Allstate Life AVD Blue Cross Life Canada Life Co-operators Life Confederation Life Equitable Life Great West Life La Capitale Assurances La Survivance L Industrielle-Alliance London Life Manulife Financial Maritime Life Metropolitan Life Mutual Life National Life North American Life Prudential of America SSQ Standard Life Sun Life Zurich Life The study compares the experience to the 1987 Basic Group LTD (87GLTD) table, the Life Waiver Study based on the CIA LTD study and a recent SOA table 2005 Group Term Life Waiver Reserve Table. The key findings from this study are; experience in Québec is significantly different than in the rest on Canada; male mortality rates are consistently higher than the female mortality rates; there is no significant difference between male and female recovery rates; there is an identifiable spike in recoveries at the time the definition of disability changes from own occupation to any occupation ; type of insurance arrangement did not have a significant impact on the termination rates; there is a lower recovery rate immediately after the claim is admitted; 3

4 there has been a slight upward trend in recovery rates over the 10-year study period ( ); and the Life waiver experience in the US is significantly different than the LTD experience in Canada. 4

5 ACKNOWLEDGEMENTS Subcommittee Members (current members) Peter Douglas (Chairperson) Don Blue Isabelle Bouchard Joel Cornberg Réjean Dion Ronald Hoskins Micheline Lafond Nathalie Laporte Cindy Lau Stéphane Levert Stella-Ann Ménard Wayne Morris Michel Simard (former members) John Have (former Chairperson) Peter Muirhead (Chairperson) Walton Achoy Karen Davies Jack Earl Denis Garand Firozali Hirji Ferdinand Iglesias Rudra Maharajh Michel Rouette Jeff Neufeld Marlene Van den Hoogen Pascal Verrette Whitman Wu Others Munich Re Outside Providers STATLAB University of Western Ontario Marc Vincelli and Bruce Jones 5

6 INTRODUCTION There is considerable interest in industry-wide studies for assumptions used in actuarial work. The Research Committee Group Life & Health Subcommittee (the Committee) understands the desire for these studies and also understands the desire of insurers to retain the information in their databases for their use. The Committee has limited the scope of this study to broad-based characteristics and intentionally leaves the more detailed studies to private studies and commercial database providers. The characteristics that are included in this study are; year, gender, region, age, duration of disability, reason for termination, cause of disability and type of insurance arrangement (Insured vs. Administrative Service Only {ASO} plans). While the information coming out of this study gives an indication of recent aggregate Canadian experience, the Committee did not analyze the data by other factors that may contribute to termination experience. These other factors could include; industry, individual s occupation, plan design, benefit level, etc. Data were collected from virtually all insurers in Canada, covering the period 1988 to The data collected represent over 300,000 records, covering over four million months of exposure and including over 140,000 terminations. This represents approximately 95% of the LTD premium in Canada. Terminations were separated between death, recovery and expiry (maximum benefit period reached). The Committee chose to compare the results of this study to three existing published tables. The first of these tables is the 1987 Basic Group LTD (87GLTD) table which is a Society of Actuaries (SOA) study developed from US data for 1975 to The 87GLTD table does not separate LTD recoveries and deaths. The second table is the Life Waiver Study based on the CIA LTD study, which does separate recoveries and deaths. The third table is the Group Term Life Waiver Table, another SOA table which does differentiate between deaths and recoveries. The graduated tables that have been produced reflect the average experience for the period and do not include any margins. Although there are several large contributors of data to the study, there is no single contributor that predominates. The largest three contributors provided 27%, 22% and 9% of the exposures. The Committee hired Statistical Laboratory (STATLAB) at the University of Western Ontario to collect and screen the data. The Committee members did not have access to the individual data records or any detailed summaries that would allow them to match experience to a particular insurance provider. The report covers; some basic characteristics of disability plans in Canada; summaries of the data; comparisons to published tables; comparisons by region, insurer, year and cause of disability; the graduation methodology; and final graduated results. 6

7 GROUP DISABILITY CHARACTERISTICS AND PLAN DESIGN Group long-term disability benefits are an important part of an employee s benefit package. The common benefits that are provided to disabled employees include; Income replacement benefits (LTD); Continuation of Life Insurance coverage (Life Waiver); and Continuation of Medical and Dental benefits. In Canada, it is typical that a common definition of disability is applied for these benefits. In the US, a different definition is often applied for the different disability benefits. When the definition of disability differs, the continuation of Life Insurance coverage often requires a much more stringent definition of disability that the LTD benefit requires. While the valuation of the obligation to provide future LTD and Life Waiver benefits to disabled employees has been a common focus, the valuation of the continuation of Medical and Dental benefits is becoming of increasing interest. Due to the uncertain future benefit level of Medical and Dental benefits these are not insurable benefits but the employer s obligation does exist. Income Replacement Benefits/Long-Term Disability (LTD) Benefits The data for this study were collected from income replacement or LTD policies. LTD has been an important group benefit in Canada for many years, and has grown significantly since Disability income periods range from two years to age 65, with the latter being most common. annual premium revenues in Canada were about $2.166 billion in The typical product provides benefits for two years, from either date of disability or end of the period, if an insured person is unable to perform the duties of his/her own occupation, ( own occupation ) and thereafter for any occupation for which the person is reasonably trained or educated ( any occupation ). A typical LTD plan provides a disabled person with 50% to 75% of pre-disability income. The coverage is coordinated with other overlapping disability coverages (such as those provided by statutory plans) so that the total benefits are not a disincentive for the disabled worker to return to work when able. The major overlapping benefits are provided either by the Canada Pension Plan (CPP) or by the Québec Pension Plan (QPP). Usually LTD contracts offset the CPP/QPP benefits entirely or partially so that claimants are limited to, at most, 85% of pre-disability earnings. Rehabilitation benefits are usually a major component of most plans. LTD plans are affected by many forces in the workplace. Over the last few decades, the Canadian workplace has experienced many changes that would impact the LTD experience, increased international competition, technology, advances in medical treatment and the aging workforce, to mention just a few. Legislative and legal changes have also impacted on the disability programs offered. On top of these uncontrollable factors, the claims adjudicators have altered their processes for rehabilitation and other return to work initiatives. These, either individually or collectively, can often explain the trends that are observed in the emerging experience. 7

8 With a few differences, the product is similar to that sold in the United States, both with reference to policy provisions and the underwriting and claims practices needed to control the results. Integration with US Social Security disability benefits parallels the integration of CPP/QPP benefits in Canada. Notwithstanding these similarities, the experience may be significantly different due to differences in the workforce demographics and other workplace factors. One significant difference between US and Canadian plans is Canadian disability plans cannot use the cause of disability to differentiate between plan provisions. In particular, mental and nervous claims cannot be treated differently than any other disability (i.e., if the disability plan provides benefits to age 65, mental and nervous claims cannot be limited to 24 months). CONFIDENTIALITY In the initial stages of the study, the Committee decided that confidentiality would be a very sensitive issue. The Committee wanted to ensure that no company would be reluctant to participate in the study for reasons having to do with confidentiality, and severely restricted access to the data submitted. 1. Only the STATLAB knew which data belong to a particular company. None of the STATLAB personnel had any links to any of the contributing companies or any of the direct insurers. The initial data gathering was delayed due to one of the senior STATLAB personnel being removed from the project so she could pursue a contract with one of the contributing companies. All other references to a particular company s data were handled by using code numbers. 2. The tables were reviewed carefully in order to ensure that no company could be identified from a particular table. For example, if only one company had been able to submit data by cause, the Committee would have excluded cause from the analysis in order to protect the identity of that company. 3. The list of participating companies deliberately excludes any reference to numbers of claims, proportion of the market held, and similar items, in order to ensure claims could not be attached to the companies. DATA COLLECTING AND REFINEMENT METHODOLOGY Period Covered: In general terms, the study covers claims active at some time during the period from the start of 1988 through the end of However, not all companies were able to provide data for this entire period. From an economic perspective, the period includes some highs and some lows - but probably does not cover a complete business cycle. From a benefit design perspective, the twoyear maximum benefit period limitation for mental and nervous conditions was removed just prior to the study period, to comply with employment equity regulations. Companies were requested to provide all the data that they had available. One company provided data back through 1979 and another back through While the Committee wanted to include as much data as possible, it was not appropriate to include data from periods that were represented by only a few companies. After a review of the pattern of companies by year, the January 1, 1988 start date was chosen. 8

9 Initial Processing of Contributions For each insurer, a profile of the data set was produced and examined. For most fields this profile consisted simply of a listing of the values that occurred in that field. These were checked against the expected values for that field, and any unexpected values were flagged for further review. For example, the gender field was expected to contain either M or F and other values were flagged. The disablement date and termination date were summarized by year and also checked against other dates submitted. The date of birth was summarized by year, and also used in conjunction with the disablement date to calculate an age at disablement. Comparisons of the report date relative to the disablement date and the date benefits commenced were done. Comparisons of the termination date relative to the disablement date, the report date and the date benefits commenced were also done. Copies of these profiles, with any concerns noted, were sent to contributing companies. Followup discussions were held as necessary to resolve any concerns. After the initial profiles were produced, additional summaries were made with appropriate follow-up as necessary. Particular note was made of net amounts that were zero and negative. A comparison was made between the date benefits commenced and the end of the period. Other and no claim termination reasons were reviewed to ensure appropriate processing. In conjunction with the termination reason, a comparison between the benefit duration information and the termination date was made. Some insurers submitted revised contributions. Other insurers submitted changes on the hardcopy error list. For each contribution, an exposure period was established with the contributor by determining a contribution start date and contribution end date. Any records with termination dates prior to the contribution start date or with disablement dates after the contribution end date were deleted. Also, on any records with a termination date after the contribution end date, the termination data were deleted (i.e., the record was counted as an active claim through the contribution end date). Note that the contribution start date was never earlier than January 1, 1988 and the end date was never later than December 31, 1997, even though companies may have submitted valid records outside this range. An exposure start date was established for each record as the earlier of the date benefits commenced and the end of the period. These fields essentially agreed on most records, but differed widely in both directions on some records. Records were deleted if the termination date was before or the same as this exposure start date. Records were also deleted if the exposure start date was after or the same as the contribution end date. Deletions were also made for certain field values being other than acceptable values or outside what came to be considered an acceptable range. A significant number of deletions were made for termination reasons of no claim, periods of greater than 25 months and unknown gender. A more moderate number of records were deleted for ages at disablement being greater than 69 and geographic regions being out of Canada. A few records were also deleted for ages at disablement being less than 15, disablement date missing and identified company specific situations. 9

10 Categorizing the Data Exposures were developed by calculating fractional durations, from date of disability- in months to exposure start dates, expiry dates and termination (decrement) dates. These were adjusted as follows: 1. Exposures were counted to end of month (or year) for decrements due to death, recovery and benefit ceasing due to change in definition; 2. Terminations due To Settlements were counted as expiries without decrements; 3. Termination reason Other was counted as a 0.75 decrement. This was coded when reason was unknown and the 0.25 reduction represents the average ratio of terminations which were due to expiries based on the whole data set; and 4. Age was calculated in quinennial age group at date of disability. DATA SUMMARY After the processing of the various data checks and corrections, there were over 300,000 records. These records provided 4,027,405 units of exposure and 140,959 terminations. Records that ceased due to the expiry of the benefit period (often age 65) were not included in the terminated record count. The 140,959 terminated records included 20,210 records that terminated due to death. The remaining 120,749 were classified as recoveries. With this classification, disabilities that terminated due to a change in definition of disability (often occurring after 24 months) were classified as recoveries. The table below provides a summary of the data and a more detailed summary is included in Appendix A. Further details are available from the CIA. Exposures Recoveries Deaths Male Female Male Female Male Female ,157 65, ,450 2,347 1,776 4, , , ,470 4,153 3,183 7, , , , ,818 6,386 4,993 11,379 1, , , , ,543 7,166 5,700 12,866 1, , , , ,732 8,324 7,234 15,558 1, , , , ,613 8,588 7,751 16,339 1, , , , ,003 7,317 6,307 13,624 1, , , , ,275 5,239 5,498 10,737 1, , , , ,330 6,224 6,694 12,918 1, , , , ,171 7,453 8,416 15,869 1, ,274 2,274,097 1,753,308 4,027,405 63,197 57, ,749 13,914 6,296 20,210 10

11 Exposures Recoveries Deaths Male Female Male Female Male Female BC 193, , ,774 4,989 5,422 10,411 1, ,640 Alberta 95,393 74, ,856 2,777 2,277 5, Saskatchewan 92,477 95, ,313 2,567 3,131 5, Manitoba 49,367 52, ,661 1,411 1,547 2, Ontario 834, ,627 1,550,879 18,234 18,691 36,925 5,167 2,907 8,074 Québec 477, , ,727 18,285 13,713 31,998 3, ,248 New Brunswick 59,236 32,114 91,350 1, , Nova Scotia 89,530 61, ,498 1,767 1,305 3, PEI 6,881 7,095 13, Newfoundland 38,897 18,514 57, , NWT 3,202 3,087 6, Yukon 2,438 2,338 4, Unknown 331, , ,895 10,683 9,553 20,236 2, ,962 2,274,095 1,753,310 4,027,405 63,194 57, ,749 13,914 6,296 20,210 Comparison to Other Tables 1987 Basic GLTD Table Most companies have used the 1987 Basic GLTD table while substituting their own experience for a specified period. In the US, the maximum period of such substitutions is two years, unless a company s experience is highly credible. In Canada, companies have often modified the 1987 Basic GLTD table or substituted their own experience for periods longer than two years. The 1987 Basic GLTD table is based on data from the period 1975 to 1980 for the first two years of termination rates, and back to 1962 for longer durations. The data are mainly US based, and do not allow for changes in demographics that underlie current claims either in the US or in Canada, or for differences in demographics and socioeconomic conditions between the two countries now or in the recent past. Nevertheless, the 1987 Basic GLTD table is the most recent and most used standard in Canada. Life Waiver Study Based on Canadian LTD Experience In November 2001, the Committee published a research paper that used the Canadian LTD experience (the predecessor of this study) to produce a disability mortality and recovery table. A LTD termination table would be the sum of these two decrements. This table was presented as an alternative to the 1970 Intercompany Group Life Waiver of Premium Table (Krieger Table). The Committee thought that creating a Group Life Waiver Study from LTD data was logical since most Group Life Waiver policies use the same definition of disability as is used in the LTD policies. Society of Actuaries 2005 Group Term Life Waiver Table In March 2006, the Society of Actuaries (SOA) published a report from their Group Life Insurance Experience Committee. They studied group life waiver claims between 1993 and 2002 inclusive. The study included 750,000 life years of experience. The study included a basic table (without margins) as well as a valuation table (with margins). The valuation table has been accepted as the valuation standard by the NAIC. 11

12 Summary of Comparisons The Canadian LTD termination experience is significantly different than what would have been expected by either the 1987 Group Long-Term Disability Table (1987GLTD) or the 2005 Society of Actuaries Group Life Waiver Table (2005 SOA Life Waiver). The Canadian LTD experience for , with some additional insurers contributing to the study is in aggregate close to the table derived from the earlier study. However, a distinct regional difference is apparent. The tables below provide a summary of the comparisons with a more detailed summary included in Appendix B. Further details are available from the CIA GLTD 1st year 2nd year 3rd year years 4 & 5 years 6 to 10 over 10 years 64,425 35,258 23,342 9,814 6,108 2, ,959 Expected 60,630 25,157 8,044 6,395 5,519 1, ,706 / Expected 106% 140% 290% 153% 111% 103% 131% LTD Termination Rates (death and recovery) (male) (female) Termination Rate (monthly) Expected (male) (87GLTD) Expected (female) (87GLTD) Duration Since Disability (months) Comments The 1987 GLTD Table does not differentiate between deaths and recoveries. The 1987 GLTD Table ignores terminations attributed to the change of definition of disability. 12

13 2005 SOA Life Waiver 1st year 2nd year 3rd year years 4 & 5 years 6 to 10 over 10 years Recoveries 59,174 29,763 20,621 7,242 3, ,749 Deaths 5,251 5,495 2,721 2,572 2,751 1,420 20,210 Terminations 64,425 35,258 23,342 9,814 6,108 2, ,959 Expected Recoveries 12,089 10,152 5,206 4,700 3, ,052 Deaths 10,618 10,991 4,780 4,977 4,373 1,721 37,460 Terminations 22,707 21,143 9,986 9,677 7,609 2,390 73,512 / Expected Recoveries 489% 293% 396% 154% 104% 89% 335% Deaths 49% 50% 57% 52% 63% 83% 54% Terminations 284% 167% 234% 101% 80% 84% 192% LTD Recovery Rates LTD Mortality Rates (male) (female) Recovery Rate (monthly) Mortality Rate (monthly) Expected (male) (2005 SOA) Expected (female) (2005 SOA) Duration Since Disability (months) Duration Since Disability (months) Comments The 2005 SOA Life Waiver Table does differentiate between deaths and recoveries. The low expected terminations and high expected deaths are consistent with a definition of disability being more stringent in the SOA study than in the Canadian LTD experience study. 13

14 CIA Table (by duration since disability) 1st year 2nd year 3rd year years 4 & 5 years 6 to 10 over 10 years Recoveries 59,174 29,763 20,621 7,242 3, ,749 Deaths 5,251 5,495 2,721 2,572 2,751 1,420 20,210 Terminations 64,425 35,258 23,342 9,814 6,108 2, ,959 Expected Recoveries 65,746 31,222 20,390 8,461 3, ,896 Deaths 5,510 5,433 2,626 2,772 3,282 1,449 21,072 Terminations 71,256 36,655 23,016 11,233 7,117 1, ,968 / Expected Recoveries 90% 95% 101% 86% 88% 245% 93% Deaths 95% 101% 104% 93% 84% 98% 96% Terminations 90% 96% 101% 87% 86% 119% 93% LTD Recovery Rates LTD Mortality Rates (male) (female) Recovery Rate (monthly) Mortality Rate (monthly) Expected (male) (CIA 88-94) Expected (female) (CIA 88-94) Duration Since Disability (months) Duration Since Disability (months) CIA Table (by region) BC Prairies Ontario Québec Maritimes Other and unknown Recoveries 10,411 13,710 36,925 31,997 7,123 20, ,749 Deaths 1,640 2,130 8,074 4,248 1,143 2,975 20,210 Terminations 12,051 15,840 44,999 36,245 8,266 23, ,959 Expected Recoveries 12,217 15,465 48,284 24,719 9,777 19, ,896 Deaths 1,892 2,297 8,257 3,980 1,749 2,897 21,072 Terminations 14,109 17,762 56,541 28,699 11,526 22, ,968 / Expected Recoveries 85% 89% 76% 129% 73% 106% 93% Deaths 87% 93% 98% 107% 65% 103% 96% Terminations 85% 89% 80% 126% 72% 105% 93% Comments The updated experience would support separating the Québec experience. 14

15 CIA Table (by year) Recoveries 4,123 7,336 11,379 12,866 15,558 16,339 13,624 10,737 12,918 15, ,749 Deaths 895 1,273 1,898 2,214 2,762 2,752 2,299 1,740 2,103 2,274 20,210 Terminations 5,018 8,609 13,277 15,080 18,320 19,091 15,923 12,477 15,021 18, ,959 Expected Recoveries 5,286 8,481 12,639 14,255 16,777 17,426 14,098 11,835 13,709 15, ,896 Deaths 952 1,441 2,063 2,266 2,693 2,784 2,320 1,869 2,189 2,495 21,072 Terminations 6,238 9,922 14,702 16,521 19,470 20,210 16,418 13,704 15,898 17, ,968 / Expected Recoveries 78% 86% 90% 90% 93% 94% 97% 91% 94% 103% 93% Deaths 94% 88% 92% 98% 103% 99% 99% 93% 96% 91% 96% Terminations 80% 87% 90% 91% 94% 94% 97% 91% 94% 101% 93% Recovery Rates Mortality Rates 120% 120% 100% 100% Ratio of Number of Recoveries to Expected from CIA Table 80% 60% 40% Ratio of Number of Deaths to Expected from CIA Table 80% 60% 40% 20% 20% 0% % Year Year Comments There is a distinctive upward trend in recovery rates. The trend in mortality rates is not as apparent. CIA Table (by gender) Expected / Expected Male Female Male Female Male Female Recoveries 63,196 57, ,749 66,477 63, ,896 95% 91% 93% Deaths 13,914 6,296 20,210 14,461 6,611 21,072 96% 95% 96% Terminations 77,110 63, ,959 80,938 70, ,968 95% 91% 93% Comments The updated experience would support the continued use of unisex recoveries with sex distinct mortality rates. 15

16 GRADUATION OF DATA General Rates were graduated from the base data with the following criteria: 1. Recovery rates were separately graduated for Québec and the Rest of Canada, and done on a unisex basis, Québec rates graduated into Rest of Canada after 36 months. 2. Mortality rates were separately graduated for Québec and the Rest of Canada, and done on a sex distinct basis, Québec rates graduated into Rest of Canada after 36 months. 3. The spike in terminations that was attributed to the change in definition of disability was removed from the base recovery rates, spikes appeared in the data at 24 months since disability, as well as after 24 months of payment; the terminations due to change in definition appeared to linger into 25 and 26 months of payment; the total impact (24 months since disability, plus 24 to 26 months of payment) was identified as the termination rate at the time the definition of disability changed. 4. A reduction in the termination rates was recognized in the first month of benefit eligibility. For instance, terminations during the sixth month of disability were much lower on plans with a sixth month period compared to plans with three or four month period. 5. Ultimate mortality rates were graduated into the Canadian Life Table (Canadian population table) between ages 76 and 85. The details of the graduation are included in Appendices C to E, Appendix C Illustrations of the results of the graduation Appendix D The graduated tables Appendix E Graduation formulas Caveats The actuary intending to use the study results will need to make allowance for several characteristics of the data: 1. The processing of the claim records initially submitted by individual companies resulted in the deletion of some claims and adjustments to others, as noted above. The Committee believes that the results have not been materially affected by such deletions and adjustments. 2. The separation of data into periods was done with ranges of days rather than exact months. 3. The termination experience is actually a blend of many different and evolving claims management practices. A company with an active and successful claims management program will expect higher termination rates than indicated by the study at early durations of disability. 16

17 However, they may find it necessary to use lower termination rates than those produced by the study beyond certain durations of disability, on the grounds that those claimants who are not managed back to work are seriously disabled and likely to remain so. 4. Participation in the study was such that each year of experience results from a different combination of participating companies. 5. The study period itself was one in which the economy experienced extreme peaks and troughs. Recognizing the strong connection between economic activity and morbidity, it may be necessary to make carefully considered adjustments to the results to reflect the expected future economic environment. 6. The typical LTD product provides for two years under an own occupation definition of disability, with an any occupation definition of disability required after two years. The graduation methodology attributed all extra terminations at the typical time the definition of disability changed from own occupation to any occupation to be caused by the change in definition. Not all plans in the study would have a two-year own occupation provision. 7. Not all companies were able to distinguish between insured and Administrative Services Only (ASO) claims. Thus, the observed raw termination rates are a blend of insured and ASO experience. On those claims where the type of administration was shown (74% of the total), termination rates on insured business were noticeably higher than on ASO business. Claims on this type of business were identified by just five companies. Therefore, caution needs to be used when interpreting the results. 8. The Committee found that the use of Other as a cause of termination was not consistent from one company to another. Counting each such termination as a 0.75 decrement (see prior section), assumes 0.25 of terminations are expiries; however, there is no reliable way to test this average assumption. 9. Since data were collected for a LTD study, the process of identifying death claims may have lacked the rigor that would have been applied if the contributors knew at the onset that the data would be used for a Group Life waiver study as well. 17

18 APPENDIX A DATA SUMMARIES Data Summary by Primary Cause of Disability and Reported Year Exposures Circulatory system diseases Mental and nervous disorders Accidents (including poisoning) Musculoskeletal system and connective tissue disorders Neoplasms Other Notprovided ,030 31,973 14,536 37,679 10,252 18,535 32, , ,185 53,578 32,430 60,228 15,728 30,249 38, , ,763 81,019 50,464 93,513 22,643 44,591 44, , ,527 92,795 56, ,401 25,608 50,134 48, , , ,305 66, ,311 32,191 78,706 49, , , ,093 62, ,319 34,837 65,273 50, , , ,172 42, ,167 28,577 51,184 48, , , ,957 31,322 85,598 25,919 41,752 20, , , ,792 31,813 95,756 30,958 49,383 28, , , ,905 33, ,866 37,700 58,139 26, , , , , , , , ,427 4,027,405 Terminations (Deaths and Recoveries) Circulatory system diseases Mental and nervous disorders Accidents (including poisoning) Musculoskeletal system and connective tissue disorders Neoplasms Other Notprovided , ,023 1,440 1,412 1, ,109 1,093 8, ,478 2,306 2,452 2,608 1,379 1,774 1,280 13, ,592 2,879 2,645 2,842 1,596 1,940 1,586 15, ,930 3,529 3,216 3,365 1,953 2,797 1,530 18, ,867 4,218 3,113 3,894 2,097 2,461 1,441 19, ,451 3,798 2,315 3,144 1,764 1,981 1,470 15, ,247 3,574 1,602 2,493 1,507 1, , ,170 4,151 1,566 2,785 1,707 1,568 2,074 15, ,524 5,764 1,935 3,644 2,206 2,058 1,012 18,143 13,921 32,407 20,888 27,275 15,688 17,845 12, ,959 Exposure by Primary Cause of Disability Terminations by Primary Cause of Disability Not provided 9.6% Circulatory system diseases 13.3% Not provided 9.2% Circulatory system diseases 9.9% Other 12.1% Other 12.7% Mental and nervous disorders 23.0% Neoplasms 6.6% Mental and nervous disorders 24.7% Neoplasms 11.1% Musculoskeletal system and connective tissue disorders 23.2% Accidents (including external causes of injury and poisoning) 10.5% Musculoskeletal system and connective tissue disorders 19.3% Accidents (including external causes of injury and poisoning) 14.8% 18

19 Data Summary by Province and Reported Year Exposures BC 14,253 17,777 29,181 32,188 39,322 40,012 37,173 35,368 55,724 76, ,774 Alberta 6,248 7,159 12,845 14,243 19,676 20,762 19,008 14,875 23,379 31, ,856 Saskatchewan 9,755 10,719 17,686 19,220 22,170 22,638 22,063 20,137 21,507 22, ,313 Manitoba 4,127 5,151 9,322 11,489 14,629 14,701 8,349 6,347 9,471 18, ,661 Ontario 74,464 89, , , , , , , , ,999 1,550,879 Québec 42,228 49,912 63,476 70,016 80,110 83,328 84,454 79,843 88,589 94, ,727 New Brunswick 4,329 4,479 7,215 7,599 11,204 11,429 10,402 10,574 12,323 11,795 91,350 Nova Scotia 7,103 8,153 12,534 14,460 19,646 20,999 20,060 15,377 16,428 16, ,498 PEI ,179 1,202 1,098 2,080 2,437 2,508 13,976 Newfoundland 3,032 3,380 3,788 4,194 5,840 6,761 6,400 7,320 8,036 8,661 57,411 Yukon ,031 1,102 6,289 NWT ,776 Unknown 8,344 77,044 93, , , ,498 37,173 9,917 6,413 3, , , , , , , , , , , ,171 4,027,405 Terminations (Deaths and Recoveries) BC ,013 1,354 1,319 1,332 1,081 1,711 2,546 12,051 Alberta ,331 5,885 Saskatchewan ,115 6,550 Manitoba ,405 Ontario 1,802 2,300 4,161 4,546 5,958 5,499 5,673 4,351 5,206 5,504 44,999 Québec 1,664 1,938 2,686 3,226 3,877 4,157 4,492 4,160 4,663 5,384 36,246 New Brunswick ,615 Nova Scotia ,648 PEI Newfoundland ,680 Yukon NWT Unknown 197 2,842 3,737 4,210 4,421 5,181 1, ,198 5,018 8,609 13,277 15,080 18,320 19,091 15,923 12,477 15,021 18, ,959 Exposure by Year Terminations by Year % % % % % % % % % % % % % % % % % % % % 19

20 Data Summary by Type of Insurance Arrangement and Reported Year Exposures ASO Insured With Refund Accounting Insured Without Refund Accounting Unknown ,685 70,783 63,916 23, , ,092 74, ,570 28, , ,079 75, ,335 40, , ,140 79, ,385 48, , ,951 92, ,206 93, , ,823 97, ,122 99, , ,930 99, , , , , , ,177 14, , , , ,009 15, , , , ,825 11, , ,662 1,062,456 1,891, ,327 4,027,405 Terminations (Deaths and Recoveries) ASO Insured With Refund Accounting Insured Without Refund Accounting Unknown ,057 1, , ,361 4, , ,920 2,466 7,309 1,582 13, ,147 2,570 8,402 1,961 15, ,423 3,122 9,439 3,336 18, ,683 3,248 9,677 3,483 19, ,510 3,337 7,051 4,025 15, ,021 4,735 6, , ,982 5,109 7, , ,500 6,048 8, ,143 17,501 35,053 69,923 18, ,959 Exposure by Type of Insurance Arrangement Terminations by Type of Insurance Arrangement Unknown 11.8% ASO 14.8% Unknown 13.1% ASO 12.4% Insured With Refund 26.4% Insured With Refund 24.9% Insured Without Refund 47.0% Insured Without Refund 49.6% 20

21 Data Summary by Type of Insurance Arrangement and Province Exposures ASO Insured With Refund Accounting Insured Without Refund Accounting Unknown BC 77, , ,032 31, ,774 Alberta 5,414 39,148 96,422 28, ,856 Saskatchewan 38,762 71,938 56,312 21, ,313 Manitoba 7,763 33,657 52,234 8, ,661 Ontario 232, , , ,508 1,550,879 Québec 80, , ,332 37, ,727 New Brunswick 5,018 31,229 43,993 11,110 91,350 Nova Scotia 15,707 64,597 60,159 11, ,498 PEI 476 8,948 3,475 1,077 13,976 Newfoundland ,182 25,746 6,698 57,411 Yukon 3 4, ,289 NWT 75 4, ,776 Unknown 132,126 7, , , , ,662 1,062,456 1,891, ,326 4,027,405 Terminations (Deaths and Recoveries) ASO Insured With Refund Accounting Insured Without Refund Accounting Unknown BC 2,460 3,497 4,850 1,244 12,051 Alberta 211 1,372 3,102 1,200 5,885 Saskatchewan 1,078 2,048 2,008 1,416 6,550 Manitoba 241 1,028 1, ,405 Ontario 5,194 12,160 20,922 6,723 44,999 Québec 2,956 11,123 20,447 1,720 36,246 New Brunswick , ,615 Nova Scotia 276 1,529 1, ,648 PEI Newfoundland ,680 Yukon NWT Unknown 4, ,614 4,382 23,198 17,501 35,053 69,923 12, ,959 Exposure by Province Terminations by Province Nova Scotia 3.8% New Brunswick 2.3% Yukon 0.1% NWT 0.2% Newfoundland 1.4% PEI 0.3% Unknown 14.3% BC 9.4% Alberta 4.2% Saskatchewan 4.7% Manitoba 2.5% Newfoundland 1.2% PEI 0.2% Nova Scotia 2.6% New Brunswick 1.9% Yukon 0.1% NWT 0.1% Unknown 16.5% BC 8.5% Alberta 4.2% Saskatchewan 4.6% Manitoba 2.4% Québec 18.3% Ontario 38.5% Québec 25.7% Ontario 31.9% 21

22 Data Summary by Elimination Period and Province Exposures 3-month 4-month 6-month 12-month Other and Unknown BC 88,410 59, ,377 40,797 83, ,774 Alberta 32,969 34,475 32,904 6,858 62, ,856 Saskatchewan 75,892 59,850 26,918 6,978 18, ,313 Manitoba 24,524 19,488 14,221 4,305 39, ,661 Ontario 318, , ,441 66, ,458 1,550,879 Québec 191, , , , , ,727 New Brunswick 23,689 22,528 16,707 9,218 19,208 91,350 Nova Scotia 51,412 37,294 19,020 8,363 35, ,498 PEI 4,839 4,752 1, ,412 13,976 Newfoundland 19,338 16,697 6,915 4,527 9,934 57,411 Yukon 3, ,044 6,289 NWT 4, ,776 Unknown 9,988 12,389 7,689 3, , , , , , ,840 1,445,603 4,027,405 Terminations (Deaths and Recoveries) 3-month 4-month 6-month 12-month Other and Unknown BC 2,934 2,541 2, ,748 12,051 Alberta 1,190 1,382 1, ,122 5,885 Saskatchewan 2,511 2,002 1, ,550 Manitoba ,262 3,405 Ontario 9,660 8,952 9,389 1,783 15,215 44,999 Québec 9,310 10,872 7,674 3,455 4,935 36,246 New Brunswick ,615 Nova Scotia 1, ,648 PEI Newfoundland ,680 Yukon NWT Unknown 1,218 1, ,233 23,198 30,548 29,712 24,096 7,275 49, ,959 Exposure by Elimination Period Terminations by Elimination Period 3-month 21.1% 3-month 21.7% Other and Unknown 35.9% Other and Unknown 35.0% 4-month 17.8% 4-month 21.1% 12-month 6.7% 6-month 18.5% 12-month 5.2% 6-month 17.1% 22

23 Data Summary by Elimination Period and Reporting Year Exposures 3-month 4-month 6-month 12-month Other and Unknown ,715 11,824 20,869 11,020 66, , ,248 16,012 26,816 14, , , ,410 37,423 48,451 21, , , ,949 47,544 53,924 22, , , ,130 62,454 63,263 25, , , ,889 70,977 65,640 26, , , ,438 75,540 64,823 25, , , , , ,672 39,771 3, , , , ,803 41,842 3, , , , ,146 42,921 3, , , , , ,840 1,445,603 4,027,405 Terminations (Deaths and Recoveries) 3-month 4-month 6-month 12-month Other and Unknown , ,668 5, , ,632 8, ,620 1,360 1, ,206 13, ,785 1,820 1, ,172 15, ,961 2,416 2, ,289 18, ,191 2,716 2, ,322 19, ,202 3,127 2, ,746 15, ,758 4,249 3,377 1, , ,934 5,676 4,228 1, , ,044 7,251 5,600 1, ,143 30,548 29,712 24,096 7,275 49, ,959 23

24 Terminations Split Between Deaths and Recoveries Terminations (Deaths and Recoveries) Recoveries Deaths , , ,336 1,273 8, ,379 1,898 13, ,866 2,214 15, ,558 2,762 18, ,339 2,752 19, ,624 2,299 15, ,737 1,740 12, ,918 2,103 15, ,869 2,274 18, ,749 20, ,959 Deaths 14.3% Recoveries Deaths BC 10,411 1,640 12,051 Alberta 5, ,885 Saskatchewan 5, ,550 Manitoba 2, ,405 Ontario 36,925 8,074 44,999 Québec 31,998 4,248 36,246 New Brunswick 2, ,615 Nova Scotia 3, ,648 PEI Newfoundland 1, ,680 Yukon NWT Unknown 20,236 2,962 23, ,749 20, ,959 Recoveries 85.7% 24

25 APPENDIX B COMPARISON TO PUBLISHED TABLES Summary of Exposures Exposures (Months of exposure for first 36 months, Annual exposures after) by gender disability Male Female 1st year 631, ,531 1,199,647 2nd year 834, ,326 1,456,018 3rd year 610, ,599 1,041,134 years 4 & 5 72,950 48, ,413 years 6 to 10 90,145 59, ,948 over 10 years 34,658 24,587 59,245 2,274,096 1,753,309 4,027,405 Exposures (Months of exposure for first 36 months, Annual exposures after) by period disability 3-month 4-month 6-month 12-month Other and Unknown 1st year 302, , , ,103 1,199,647 2nd year 273, , , , ,535 1,456,018 3rd year 196, , , , ,690 1,041,134 years 4 & 5 24,643 17,045 24,293 13,365 42, ,413 years 6 to 10 34,732 17,978 31,186 15,043 51, ,948 over 10 years 16,700 4,982 12,406 4,958 20,199 59, , , , ,840 1,445,603 4,027,405 Exposures (Months of exposure for first 36 months, Annual exposures after) -- by region disability BC Prairies Ontario Québec Maritimes 25 Other/ Unknown 1st year 107, , , ,567 85, ,408 1,199,647 2nd year 140, , , , , ,534 1,456,018 3rd year 99, , , ,774 85, ,680 1,041,134 years 4 & 5 11,735 12,937 50,667 21,081 10,544 14, ,413 years 6 to 10 13,840 16,249 64,302 25,778 13,437 16, ,948 over 10 years 5,002 5,549 26,230 11,390 5,527 5,547 59, , ,830 1,550, , , ,960 4,027,405 Exposures (Months of exposure for first 36 months, Annual exposures after) by reporting year disability st year 52,686 85, , , , , , , , ,322 1,199,647 2nd year 63, , , , , , , , , ,461 1,456,018 3rd year 45,825 69,634 99, , , , ,324 92, , ,423 1,041,134 years 4 & 5 5,386 7,604 11,069 12,070 14,483 15,861 14,289 11,787 13,451 15, ,413 years 6 to 10 5,996 8,805 12,987 14,317 17,301 18,548 16,793 14,547 18,357 22, ,948 over 10 years 1,941 2,700 4,179 4,829 6,237 7,452 7,610 6,080 7,962 10,255 59, , , , , , , , , , ,171 4,027,405 Exposures (Months of exposure for first 36 months, Annual exposures after) by age at disability disability under st year 18,789 62, , , , , , , ,864 1,199,647 2nd year 16,530 59, , , , , , , ,027 1,456,018 3rd year 9,479 34,694 69,201 96, , , , , ,029 1,041,134 years 4 & ,532 7,049 10,646 13,546 17,640 24,399 31,918 11, ,412 years 6 to 10 1,394 4,832 8,955 13,972 18,808 25,923 39,046 35,183 1, ,947 over 10 years 59,247 59,247 47, , , , , , , , ,464 59,247 4,027, and over all ages

26 Summary of Recoveries Recoveries by gender disability Male Female 1st year 30,311 28,863 59,174 2nd year 15,777 13,986 29,763 3rd year 10,421 10,200 20,621 years 4 & 5 4,202 3,040 7,242 years 6 to 10 2,123 1,234 3,357 over 10 years ,196 57, ,749 Recoveries by period 3-month disability 4-month 6-month 12-month Other and Unknown 1st year 15,121 15,712 8, ,932 59,174 2nd year 4,673 5,662 5,927 2,631 10,870 29,763 3rd year 4,473 3,435 3,821 1,416 7,476 20,621 years 4 & 5 1, ,305 1,121 2,573 7,242 years 6 to ,221 3,357 over 10 years ,174 26,235 20,380 5,666 42, ,749 Recoveries by region disability BC Prairies Ontario Québec Maritimes Other/ Unknown 1st year 4,556 6,364 15,430 19,108 2,880 10,836 59,174 2nd year 2,573 3,178 9,967 7,524 1,763 4,758 29,763 3rd year 2,144 2,873 7,110 3,247 1,464 3,783 20,621 years 4 & ,756 1, ,242 years 6 to , ,357 over 10 years ,411 13,710 36,925 31,998 7,123 20, ,749 Recoveries by reporting year disability st year 2,041 3,834 5,940 6,560 7,625 7,986 6,539 5,199 6,118 7,332 59,174 2nd year 1,040 1,752 2,662 3,142 3,942 3,899 3,449 2,567 3,270 4,040 29,763 3rd year 694 1,238 1,826 2,091 2,671 3,025 2,269 1,828 2,141 2,838 20,621 years 4 & ,242 years 6 to ,357 over 10 years ,123 7,336 11,379 12,866 15,558 16,339 13,624 10,737 12,918 15, ,749 Recoveries by age at disability disability under st year 1,900 5,316 8,462 9,598 9,546 8,879 7,303 5,424 2,747 59,175 2nd year 745 2,310 4,056 4,856 4,893 4,740 4,036 2,957 1,170 29,763 3rd year 454 1,448 2,670 2,984 3,261 3,195 3,139 2, ,623 years 4 & ,004 1,082 1,089 1,213 1, ,242 years 6 to ,358 over 10 years ,304 9,770 16,303 18,853 19,209 18,390 16,421 12,732 5, , and over all ages 26

27 Summary of Deaths Deaths by gender disability Male Female 1st year 3,425 1,826 5,251 2nd year 3,722 1,773 5,495 3rd year 1, ,721 years 4 & 5 1, ,572 years 6 to 10 1, ,751 over 10 years 1, ,420 13,914 6,296 20,210 Deaths by period 3-month disability 4-month 6-month 12-month Other and Unknown 1st year 1,220 1, ,936 5,251 2nd year , ,967 5,495 3rd year ,721 years 4 & ,572 years 6 to ,751 over 10 years ,420 4,374 3,477 3,716 1,609 7,034 20,210 Deaths by region disability BC Prairies Ontario Québec Maritimes Other/ Unknown 1st year ,050 1, ,251 2nd year ,226 1, ,495 3rd year , ,721 years 4 & ,572 years 6 to , ,751 over 10 years ,420 1,640 2,130 8,074 4,248 1,143 2,975 20,210 Deaths by reporting year disability st year ,251 2nd year ,495 3rd year ,721 years 4 & ,572 years 6 to ,751 over 10 years , ,273 1,898 2,214 2,762 2,752 2,299 1,740 2,103 2,274 20,210 Deaths by age at disability under disability st year ,040 1, ,251 2nd year ,005 1, ,495 3rd year ,721 years 4 & ,572 years 6 to ,751 over 10 years 1,420 1, ,397 1,938 2,909 3,905 4,769 2,570 1,420 20, and over all ages 27