Canadian Institute of Actuaries Institut Canadien des Actuaires MEMORANDUM

Transcription

1 Canadian Institute of Actuaries Institut Canadien des Actuaires MEMORANDUM TO: FROM: All Members and Students of the CIA practising in the area of pension plans Douglas P. Chandler, Chairperson of the Task Force on Commutation of Pensions Due to Shortened Life Expectancy DATE: April 23, 2001 SUBJECT: Discussion Draft of the Standard of Practice for the Computation of Pension Benefits Due to Reduced Life Expectancy COMMENT DEADLINE: May 31, 2001 Document The Task Force on Commutation of Pensions Due to Shortened Life Expectancy has prepared the attached Discussion Draft of the Standard of Practice for the Computation of the Commuted Value of Pension Benefits Due To Reduced Life Expectancy (the draft). This memorandum summarizes the considerations of the Task Force on Commutation of Pensions Due to Shortened Life Expectancy in preparing this discussion draft. Background As a result of the implementation of Regulation 144 in Ontario, former members of a pension plan, including retirees whose pensions have already started, have the option to commute their pensions if they have an illness or physical disability that is likely to shorten their life expectancy to less than two years and have a medical certification to that effect. For this purpose, the commuted value is to be calculated using methods and actuarial assumptions that are consistent with accepted actuarial practice. The Pension Plan Financial Reporting (PPFR) Committee previously concluded that the current standards of practice do not cover this situation. An educational note to this effect was published by PPFR in July The task force has been established to determine accepted actuarial practice and draft an appropriate standard for these calculations. The CIA and others have suggested changes to the Regulations that would remove the commutation option for individuals whose pensions have already started. If the government accepts those suggestions, a new standard may not be required. However, that outcome is not a certainty. Meanwhile, some pensioners have elected the commuted value option, so actuaries have an urgent need for guidance in this area. The task force is also of the view that the regulation, in its current form, is ill advised, in respect of pensions in payment. Secretariat: Albert St., Ottawa, ON K1R 7X7 (613) FAX: (613)

2 In particular: Medical opinions as to life expectancy can never have certainty. In some cases, individuals may commute their pensions and still live for considerable time beyond the assumption underlying the commuted value, having given up their regular pension income. Even though spousal consent is required, commutation may leave the surviving spouse without the regular pension income that would have been available absent the commutation Plan costs will likely be increased even with the use of a reduced life expectancy assumption. Retirees will have a financial incentive to request commutation when they are very close to death and thereby receive payments reflecting a period beyond the time the pension would have been paid in the normal course. The attached table illustrates the percentage increase in annuity values to reflect an additional payment at the end of life for alternative assumed survival periods. Nevertheless, pending any change in the regulation, it is necessary to proceed with a Standard of Practice on which actuaries can rely. This standard has been developed in the specific context of the Ontario Pension Benefits Act requirement. They are applicable in other directly comparable situations, but the task force has not contemplated other applications where different medical certification requirements might apply. General Principles The draft reflects the principle that, for a former member entitled to a deferred pension, the lump sum value should reflect the present value of death benefits that would arise on the former member s death. The task force concluded that it would be a satisfactory approximation to use the present value of the benefits payable if the former member had died at the date of calculation for this purpose. Interpretation of Medical Opinions The task force considered whether the commuted value in a particular case should vary based on the specific terms of the medical certificate or other medical evidence. Considerations include: If a medical certificate indicates that life expectancy is reduced below a specific period lower than two years - for example, 12 months - should the calculation reflect a shorter survival period than where life expectancy less than 24 months is indicated? The task force concluded this would not be appropriate since the regulation only requires certification of life expectancy less than 24 months. Further, it is anticipated that, where life expectancy less than 24 months is indicated, it is likely that the average survival period will be significantly lower than 24 months in any event. If a medical certificate is provided that suggests a life expectancy greater than the survival period used in our standard but less than 24 months, should the commuted value be increased? The task force concluded this would be inappropriate. Again, medical certifications are necessarily uncertain. The number of cases where a physician is likely to opine on this basis is likely to be extremely small. Making adjustments would then require considerable judgement and would open the door to possible abuse. 2

3 Should the calculation be adjusted on the basis of specific medical evidence provided in a particular case? For example, for particular conditions, such as cancer, could an algorithm be established based on specific diagnostic test results? The task force concluded that this would not be practical. Apart from the difficulties of developing such a procedure, it could not possibly cover all possible conditions and, importantly, plan administrators have no authority to collect such data. The task force, therefore, concluded a single uniform methodology should be used in all cases based only on certification of less than 24 months life expectancy. The task force contacted the Ontario College of Physicians and Surgeons, who referred the task force to the Ontario Medical Association in order to seek the views of the medical profession on the interpretation of medical certificates and the Ontario regulatory requirement. At the date of writing, no response to these enquiries has been forthcoming. Calculation Date The task force concluded that the calculation date should be based on the date of the medical certification as the most objective basis to ensure that calculations are made, as far as possible, on a consistent basis. Other options, such as the date of receipt of the commutation request and supporting documents by the plan administrator, or the later of the date of the former member s application and the date of the medical certificate, were rejected. Selection of Survival Period Medical practitioners are asked to opine on shortened life expectancy for a variety of reasons, including: eligibility for hospice admission; eligibility for experimental drug studies; appropriateness of intrusive, painful, and perhaps futile therapies; guidance to patients in organizing their personal affairs; payment of living benefits under life insurance policies; and physician-assisted suicide (at least hypothetically). Because of the importance of these opinions to individual health outcomes and to the overall cost of health care, they have been the subject of numerous research studies. The task force reviewed medical literature evaluating the accuracy of practitioners prognostications. The evidence indicates that: medical practitioners interpret life expectancy less than x months to mean the patient is very unlikely to survive for more than x months; medical practitioners are optimistic in their estimates of life expectancy, even when the consequence of overestimating life expectancy is unfavourable for their patient; medical practitioners believe that optimism can modify disease, and so are even more optimistic in their communications to patients than in their own estimates; and subjective estimates of life expectancy based on a medical practitioner s personal knowledge are not as accurate as statistical estimation based on specific measurable conditions. 3

4 Canadian life insurance companies pay living benefits based on a physician s opinion of shortened life expectancy, but the advance is usually much less than the total death benefit, and the physician s opinion is often supported by the insurer s documentation of the waiver of premium benefit. Consequently, thorough underwriting or accurate forecasting does not seem to be a major issue. Viatical companies will purchase life insurance policies from terminally ill policyholders. These contracts are permitted in some Canadian provinces, and were recently permitted in Ontario. Viatical contracts are uncommon in Canada but more common in the United States. U.S. minimum amounts are quite high, and, consequently, underwriting is more intense, depending on independent testing for specific conditions rather than physicians opinions. This underwriting is apparently financially successful. However, viatical companies are now offering seniors benefits with no or less medical evidence, and we understand they are encountering significant anti-selection. The task force was not able to locate actuarial studies on mortality after payment of living benefits or viatical settlements. However, the studies would be of very limited relevance, given the differences in underwriting and financial motivations. Given the limited scientific data available, some considerations of the task force are: If there were no other influences involved, one might expect a survival period of about eight to 12 months. For example, one line of thinking is that former members who meet the certification requirement will experience a uniform distribution of deaths over two years. This leads to an average life expectancy among such members of 12 months. A constant force of mortality with a 95% probability of death within two years would result in an average survival period of eight months. Former members in receipt of pension have an incentive to maximize their total benefits by collecting as much pension as possible as a regular pensioner before cashing in their remaining commuted values. Many conditions cannot be diagnosed until they are in a terminal stage with life expectancy less than one year. Former members eligible for such commuted values are likely to demonstrate a considerable degree of inertia until death is imminent and the application for a commuted value cannot be further delayed without considerable risk. If and when statistics become available, the remaining lifetimes of such members after submission of their applications is likely to be very low perhaps in the range of three months. While some might expect that doctors will report in a manner that lengthens the estimated life expectancy, the more common view seems to be that doctors will be reluctant to state that life expectancy has been reduced to less than two years until life expectancy is in fact very short. Furthermore, it is expected that doctors will be reluctant to opine about any specific life expectancy, other than that life expectancy has been reduced to less than two years. Experience is likely to be influenced by the assumption itself. A longer assumed survival period will provide a higher benefit which will be more attractive to more former members who may not be extremely close to death. A shorter survival period assumption will provide less incentive to apply until the expected survival period is very short, resulting in a shorter average survival period. As reflected in the Discussion Paper, the task force has not come to a final conclusion regarding the appropriate survival period assumption. However, based on the analysis to date, an assumed fourmonth period is likely to be the recommendation. 4

5 Use of Fixed Term Equal to Expected Life Expectancy The task force recognizes that, generally, it is inappropriate to value life annuity benefits using a term certain annuity equal to the expected survival period. However, for the purpose of this particular calculation, the resulting difference will be immaterial, considering: the impracticality of creating experience mortality rates within the 24-month period; the uncertain nature of any assumption regarding life expectancy in this situation; and the desirability of avoiding more complex calculation methodologies unless a clearly better result can be obtained. TABLE Illustrations of the impact of Shortened Life Expectancy Payments: % Increase in Annuity Factor for Additional Payment at Death of Age Sex 4 Months 8 Months 12 Months % % % 55 M M M M M F F F F F Basis: GAM 1983 mortality, 7% interest, life annuity 5