This article was first published in the third-quarter 2013 edition of Personal Finance magazine.
One of the big problems in planning for retirement is that none of us has much idea of how long we will live.
Life assurance companies selling guaranteed annuities (pensions) use the law of averages to decide how long they will need to keep paying you a pension. They have records (mortality tables) that show how long people are expected to live in retirement – currently 82 or 83 on average for women and 78 or 79 for men who retire at age 65 – but even these are not entirely accurate (see “Mortality measures in practice”, below). If you outlive the expected average to, say, age 90, the extra money you receive as a guaranteed pension for those additional years comes from people who die before the average age.
However, the average age presents a problem if you are using or planning to use an investment-linked living annuity (illa) to fund your retirement. Unfortunately, in the recent past, some financial advisers based their calculations on the average ages when deciding on how long an illa should last, not taking account of the fact that half of all illa pensioners will live longer than the average.
Illa pensioners can, on average, expect to live longer than those on life assurance guaranteed pensions.
Unlike those with guaranteed annuities, there is no cross-subsidisation for illa pensioners if you outlive the average life-span. This, combined with other factors, such as a high drawdown rate and unwise higher-risk investments, has resulted in many illa pensioners becoming impoverished in their lifetimes.
When you take out risk life assurance, life assurance companies price your premiums according to the risk they assume they will be taking, so they ask you all sorts of questions about the state of your health, your age, your job and your hobbies, where you live ... all the way through to your level of education.
There is currently no such assessment of the probability of how long you will live in retirement when you make the choice between a guaranteed annuity and an illa. Until now, that is.
Research on “real-age” life expectancy (which takes account of lifestyle and other factors that affect life expectancy) and the consequent probability of illa pensioners surviving financially has been undertaken by actuary Professor Niel Krige, who is in charge of the postgraduate diploma course in financial planning at the University of Stellenbosch. As a former managing director and deputy chairman of Momentum Life and a non-executive board member of financial services company MMI Holdings and its subsidiary, MMI Investments, Krige knows his way around pension products.
He presented his research at his recent inaugural lecture as professor, saying that predicting life expectancy is very important for two reasons:
* Given the diverse nature of the South African population, the national average life expectancy cannot be applied to everyone; and
* The calculation of retirement duration forms a vital part of the retirement planning process.
With more information about their life expectancy, he says, illa pensioners can make more informed investment decisions, taking account of the factors that will affect their life-span, which, in turn, affect the probability of having sufficient funds during retirement.
Krige’s research is based on two models: one to better predict life-span; the other to match that anticipated life-span to optimal illa drawdown rates within the allowed range of between 2.5 percent and 17.5 percent of the annual capital value.
There can be no 100-percent guarantee that an illa will not suffer “capital death” before you die, Krige says. However, the two models, while not a guarantee of success, will help to give illa pensioners a lot more peace of mind than many currently have.
Krige’s research shows that “real age” – which is calculated using probabilities related to lifestyle and other factors – rather than “calendar age” should be used when planning for retirement. His real age model incorporates factors such as gender, province of residence, income, HIV status, ethnic background, weight, exercise, family illness history, stress, substance abuse and diet.
So, for example, Krige’s model shows that 65-year-old, male high-income earners with healthy lifestyles who live in the Western Cape may have a real-age adjusted life expectancy in retirement that is more than double that of the country as a whole.
Krige says the research results should not be used only in retirement. “The model has significant financial implications and should be factored into retirement planning at an early age,” he says.
In other words, if you are in the group that can expect to live for many years in retirement, you will need to save a lot more retirement capital than someone who expects to live for a shorter time.
Conversely, it is prudent to err on the conservative side if your real-age-adjusted life expectancy is shorter than your life expectancy based on normal mortality tables. But even if you expect to have a shortened life-span, you cannot spend more in retirement or save less for retirement, in case you are the exception and live longer than expected.
Krige’s second model, on financial survivability, deals with the question of how long a given amount of capital will be able to fund a pensioner if the following five parameters are known:
* Expected retirement duration (years between the date of retirement and the expected date of death);
* Investment returns;
* Inflation;
* The annual withdrawal amount; and
* The initial capital amount available.
Krige says an important fact revealed by his research is that an illa requires a meaningful investment in equities if you are to be protected against the negative effects of inflation.
Although equities have significantly out-performed bonds and cash over longer periods, the volatility of equity returns is also substantially higher. The danger of volatility is that a pensioner may have to cash in a significant number of equities when equity prices are low, resulting in limited upside potential when equity prices recover.
Krige warns: “This may result in an increased risk of money death.” The terms “money death” or “capital death” refer to the risk that an illa pensioner may run out of money during retirement.
The implication of this research is that retirees have to accept that there are very few scenarios where their financial survival probability is close to 100 percent. The most important precautionary measure available to illa pensioners is to start with the lowest possible withdrawal rate of 2.5 percent at retirement. This rate can be increased at a later stage should post-retirement investment performance be above average.
If you cannot start on the minimum withdrawal rate, you should postpone your retirement. Krige says delaying retirement has the threefold advantage of increasing the funds available for retirement due to investment performance, adding further contributions to your savings, and shortening the period in retirement for which provision has to be made.
MORTALITY MEASURES IN PRACTICE
There is no single measure provided by the keepers of mortality tables that you can use as a guideline to gauge, even approximately, how long you may live.
There are general figures – for example, those you find on the website of the World Health Organisation (WHO). And even then there is no single figure.
Most mortality tables – statistics on how long people are expected to live, on average – are divided into male and female tables, because women generally live longer than men. Then they are usually differentiated by age.
So the WHO provides a life expectancy for South Africans at birth of 50.1 years for baby boys and 48.3 years for baby girls. For boys and girls who live to age 10, the situation changes in favour of girls, who are then expected to live longer than boys, namely 58.3 years as opposed to 58.2 for boys. The advantage for girls then remains for life.
If you survive to age 60, the WHO life expectancy for men is 75.1 years and women 77.8. And if you reach 65, you can expect to live, on average, to 77.5 as a man and 79.6 as a woman.
The reason for the extended life expectancy when you reach your sixties is that people who have died from childhood diseases and as a consequence of the HIV/Aids epidemic are now out of the calculations. The WHO estimates that 44.07 percent of all South African deaths are due to HIV/Aids, with influenza and pneumonia coming second at 8.88 percent of deaths.
Violent deaths come eighth at 2.27 percent and road accidents eleventh at 1.45 percent of deaths.
The longer you live, the longer you can expect to live, as you start escaping the high-risk ages where diseases such as HIV/Aids strike.
But the WHO life expectancy figures, in the seventies for those who reach age 65, seem to be low when compared with the expected longevity figures used for those buying financial services products.
For example, Premiums & Problems, the Old Mutual guidebook for financial planners, puts the anticipated average life span of 65-year-olds buying voluntary purchase annuities (pensions bought with discretionary money such as savings) at 78.81 for men and 82.09 for women.
Even these figures are lower than the generally accepted life expectancy figures of people who use investment-linked living annuities as a pension source: 83 for men aged 65 and 87 for women aged 65.
Actuary Brett Cameron, head of Old Mutual retail underwriting and claims, says most financial services companies tend to use the experience of their own customer base to calculate their life expectancy figures.
So the WHO figures would not be used by Old Mutual to calculate the average life expectancies of men and women of different ages.
The company compiles figures based on the actual average life spans of its own customers. And in doing so it often calculates the life expectancies of different cohorts. For example, it will have life expectancy figures for people buying life assurance, other figures for people buying guaranteed pensions, and even figures for different types of guaranteed pensions, Cameron says.
Calculating longevity is very important, both for financial services companies and for you. If you do not have a rough estimate of how long you can expect to live, you cannot get anywhere near estimating things such as how much money you may need for retirement and how long or even how much life assurance you may need.
Financial services companies, particularly life assurance companies, need to know – particularly if they have provided you with some sort of guarantee – how long they will need to keep paying you and when they can expect to pay out a death benefit.
Actuary John Anderson, managing director of research and product development at Alexander Forbes Financial Services, says that his company’s longevity figures are based on a house view that, in turn, is based on extensive research done over a number of years that take more than age and gender into consideration. The calculations include income levels and anticipated future improvements in mortality as a result of things such as improved health.
The issues of improved health and income are linked, because people with higher incomes can afford a more healthy lifestyle and have access to better health care.
So, on Alexander Forbes’s mortality tables for pensioners, a 65-year-old man with an income of up to R20 000 a year can expect to live, on average, to age 80.1; between R20 000 and R40 000 to age 80.5; between R40 000 and R60 000 to age 81.7; and over R60 000 to age 83. A 65-year-old woman with an income of up to R20 000 a year can expect to live to 83.6; between R20 000 and R40 000 to age 84.4; between R40 000 and R60 000 to age 85.8; and above R60 000 to age 87.3.
Anderson says Alexander Forbes believes the figures represent reasonable estimates for people who participate in formal retirement arrangements (via pension funds, provident funds or retirement annuities). Importantly, the figures allow for future improvements in mortality and are not just reflective of the current mortality rates.
He says the calculations include employed individuals in retirement arrangements only and not unemployed South Africans. This would explain the difference between the Alexander Forbes life-span figures for the lower income bands and the general WHO figures.
FACTORS THAT INFLUENCE YOUR LIFESPAN
If you live in the Western Cape, you have a far better chance of a long retirement than if you live in the Free State – and your chances increase further if you are female.
Province of residence and gender are the two base factors that were used by Professor Niel Krige of the University of Stellenbosch in his model to calculate “real age” rather than “calendar age” for life expectancy in retirement.
Krige’s real-age model does not take account of the fact that a person’s life expectancy increases with an increase in age. The basis of the model is life expectancy according to province and gender. Years are then added or subtracted to take account of other factors. These factors are:
Living standards measure (LSM) grouping
LSM is the most widely used market segmentation tool in South Africa, cutting across race, gender and age. It groups people into 10 categories according to their living standards. Given South Africa’s high Gini coefficient (an international measure of income inequality), only a relatively small portion of the population has a life expectancy that would be found in a developed country.
The difference in life expectancy between the high and low LSM groups is 30 years, and this forms the basis of the LSM input into the real-age model.
HIV status
HIV/Aids has an important role in the model because the disease is an epidemic in South Africa and is unlikely to be eradicated in the near future.
According to the National Health Service (NHS) in the United Kingdom, a person with HIV/Aids can live to the age of 66 years, if treated correctly. The NHS study states that this is 20 years lower than the normal life expectancy in the UK. So if a person has HIV/Aids, Krige’s model assumes that his or her life expectancy will decrease by about 25 percent.
Ethnic background
Racial profiling, although banned for political reasons, still has an important role to play in determining life expectancy in South Africa.
Although income and education levels are much more significant determinants of life expectancy, Krige says a study conducted in 2005 showed that the life expectancy of black people is six years less than that of white people. For the purposes of the model, the assumption is that the life expectancy of white people increases by three years. No adjustment is made for coloured and Indian people, whereas the life expectancy of black people is reduced by three years.
Weight
The World Health Organisation states that obesity is the fifth biggest cause of death worldwide. It is also one of the first factors considered in any real-age survey. Obesity is becoming a problem in South Africa, despite it still being a developing country, with 59 percent of women and 29 percent of men considered to be obese. A study in the United States found that the life expectancy of 40-year-old non-smoking men who were 20kg overweight was eight years lower than their counterparts with optimal body mass.
Krige says that, theoretically, each person’s body mass index should be calculated, but for simplicity, “kilograms overweight” was used. The relationship between kilograms and decrease in life expectancy is assumed to be 0.4. So, if a person is 10kg overweight, life expectancy is reduced by four years.
Exercise
A 2011 study concluded that exercising for 30 minutes a day increases life expectancy by three to four years. However, there is a limit to which daily exercise will add years to your life. Exercising for 10 hours a day will obviously not increase life expectancy by 60 to 80 years, Krige says. A cap of five years is set in the model if a person exercises for 90 minutes or more per day.
Family history of illness
Hereditary diseases are not uncommon, and clearly a person is at higher risk of a reduced life expectancy if the gene for one of these diseases is in the family.
Krige says it is almost impossible to determine the relevance of a family history of illness for each person, so a notional value was applied in the model. According to the Bankrate Life Expectancy Calculator (2012), life expectancy decreases by three years if both parents die of illnesses before the age of 60. This is the extreme case, so for the purposes of the model, family history is broken down into three categories: high prevalence, prevalent and not prevalent. In each case, three years, one-and-a-half years and zero years are subtracted from life expectancy, respectively.
Stress
Krige says life has definitely become more stressful, faster-paced and less forgiving than it was 50 years ago. Stress can have negative effects on your health and thus diminish life expectancy. Two years is the maximum decrease in life expectancy built into the model.
Substance abuse
Smoking, drugs and (prescription) substance usage can cause long-term damage. The model reduces life expectancy by four years for smokers and by two years for people who smoked at any time during their lives. Drugs and substance abuse may have even more serious implications, but since this information is not often shared with a financial planner, it is not taken into account for the purposes of this model.
Diet
Two years are added to your life expectancy if you eat healthily. No adjustment is made for an average diet, and two years are subtracted for an unhealthy diet. A balanced meal is considered to be meat (in moderation), salads and/or vegetables, and at least three meals a day must be consumed. Fast foods and high-starch meals do not fall into the healthy diet category.
THE RISKS YOU FACE WITH ILLAs
The higher the drawdown rate of an investment-linked living annuity (illa) and the longer your anticipated years in retirement, the less likely you are to survive financially, even if you opt for a high-risk (75 percent in equities) investment portfolio.
Niel Krige of the University of Stellenbosch says, however, that if you opt for a higher drawdown (withdrawal) rate and have an extended retirement, you will be better off in a high-risk portfolio than in a low-risk portfolio (25 percent in equities).
Conversely, with a low initial drawdown rate and a relatively short expected retirement period, your financial survival probability is higher in the case of a low-risk portfolio. (Financial survival probability is the probability of having enough capital to maintain a desired withdrawal rate for the expected duration of your retirement.)
For example, if your initial drawdown rate is 2.5 percent and you expect to live to 80, the probability you will survive financially if you have a low-risk portfolio is 83 percent compared with 77 percent if you have a high-risk portfolio.
So if you want a high drawdown rate and face a long retirement period, you have to accept higher risks, and thus higher volatility, in order to achieve higher returns, Krige says.
There are five main factors that will determine whether you, as an illa pensioner, will outlive your capital in retirement or will be financially secure until you die. They are: retirement duration (that is, the number of years between retirement date and death), investment returns, inflation, annual withdrawal amount and initial capital amount.
Krige’s pension drawdown model assists in retirement planning by showing how retirement duration and withdrawal rates change your financial survival probability.
Retirement duration should be based on Krige’s “real-age” adjusted life expectancy model. By using this model, illa pensioners can decide whether they are comfortable with the financial survival probability of their chosen drawdown rate, or whether they should reduce their drawdown rate to increase their financial survival probability.
To calculate what the financial survival probability is for a given age in retirement, Krige says the actual annual pension as a percentage of the retirement capital available is compared with the desired pension as specified by the living annuitant for each possible return scenario.
Initial benefits are determined as a fixed percentage of the capital amount available, with withdrawals increasing in line with inflation at six percent a year. The financial survival probability is the number of hypothetical outcomes in which the actual benefit received was at least equal to the desired benefit, divided by the total number of iterations (10 000 possible outcomes, mainly determined by different market conditions).
The research results of the pension drawdown model are presented graphically for each of three investment risk profiles: high risk (75 percent of the investment portfolio is in equities), medium risk (50 percent equities) and low risk (25 percent equities), showing clearly that the financial survival probability depends on the drawdown rate and the pensioner’s age. All three scenarios assume a retirement age of 65.
So in Graph 1 (the link to the graphs is at the end of the article), the financial survival probability of someone retiring at age 65 with an initial drawdown rate of five percent, in the case of a low-risk portfolio is 80 percent at age 72, 70 percent at age 74, 60 percent at age 79, 50 percent at age 90 and about 40 percent at age 100. Increase the drawdown rate to 12.5 percent and the financial survival probability is about 10 percent at any age in retirement. It is only with an initial drawdown rate of 2.5 percent and a short-lived retirement that you come near to any certainty that you will not outlive your retirement capital.
In Graph 2, the financial survival probability of someone retiring at age 65 with an initial drawdown rate of five percent in the case of a medium-risk portfolio is 80 percent at age 70, 70 percent at age 74, 60 percent at age 83 and less than 50 percent at age 100. Increase the drawdown rate to 12.5 percent, and the financial survival probability increases from the 10 percent in the case of the low-risk portfolio to 20 percent for any age group. Again, a drawdown rate of 2.5 percent and a short retirement are the only real protection against outliving your retirement capital.
In Graph 3, the financial survival probability of someone retiring at age 65 with an initial drawdown rate of five percent in the case of a high-risk portfolio results in a financial survival probability of 80 percent at age 69, 70 percent at age 74, 60 percent at age 86 and about 50 percent at age 100. Increase the drawdown rate to 12.5 percent and the financial survival probability increases from the 10 percent of the low-risk portfolio and the 20 percent of the medium-risk portfolio to almost 30 percent at any age. Again, only a withdrawal rate of 2.5 percent and a short-lived retirement can give you any certainty of outliving your capital.
Krige says all three scenarios show that a common financial survival probability exists regardless of whether a low-risk portfolio or a high-risk portfolio is selected.
In Graph 4, three financial survival probabilities (50, 70 and 90 percent) are plotted for the three risk portfolios. The graph shows that a financial survival probability of 70 percent produces the same line for all three. This means that when an illa pensioner chooses a financial survival probability of 70 percent, it does not matter in which portfolio he or she invests. Therefore, it would make perfect sense to opt for the low-risk portfolio.
If you are looking at a financial survival probability of below 70 percent, you should invest in a high-risk portfolio, again irrespective of your age or withdrawal amount. The converse is true for higher financial survival probabilities.
If you are looking at a financial survival probability of above 70 percent, a low-risk portfolio should be chosen. The terminal age and withdrawal amount are independent, and the portfolio could be chosen based solely on the financial survival probability.
“This unexpectedly simple result is an elegant step for retirement planning, in that it should ensure that an illa pensioner understands the concept of financial survival probability and that the decision on an initial withdrawal rate and risk profile is a more informed one,” Krige says.