Mountain Vista Winter Summit View. ProPhoto RGB.

This article was first published in the second-quarter 2012 editon of Personal Finance magazine

Assuming that the returns on your investments will remain the same, on average, year after year can play havoc with how much money you expect to have when you reach retirement age and with calculations on whether your accumulated savings will maintain your pension until you die if you purchase an investment-linked living annuity (illa).

This warning comes from one of South Africa’s leading retirement investment researchers, Daniel Wessels, a financial adviser at Martin Eksteen Jordaan Wessels.

Yet more often than not, investors, their financial advisers and product providers use what is called the “constant return assumption” to calculate whether the investor is on target to have

sufficient capital at retirement and to decide how much an illa annuitant can withdraw as a pension in retirement.

The constant return assumption means the market cycles and volatility are smoothed out, normally based on an average annual return over the longer-term performance of the different asset classes of cash, property, bonds and shares.

In simple terms, this means adding up all the returns over, say, 20 years, dividing the total by 20 to get an average return, and then assuming that this will be the annual return into the future, with each year providing the constant return.

Wessels says the use of the constant return assumption is a fundamental flaw in conventional retirement planning.

“Returns from financial markets reflect two key characteristics, namely mean reversion and volatility. This implies that future returns must, to a certain extent, be unpredictable when planning ahead.

“Our future returns might be very different from what we experienced in the recent past.”

Wessels says if you want to be sure your retirement savings will not expire before you do, you will have to be a lot more conservative than most people are in their retirement planning.

You need to take account of what he calls non-conforming or “out-of-the-norm” returns, or worst-case scenarios.

For example, if you had retired at the end of 1988, you would have retired with a lot less money than if you had retired in 2008, because the investment markets provided far better returns between 1998 and 2008 than in the previous decade.

If two people saved at the same level in exactly the same underlying investments for the same length of time, the one who retired in 2008 would have had retirement savings equal to 14 times his or her final annual pensionable income, but the 1998 retiree would have had savings equal to only eight times his or her final income.

The problem is that most advisers would have assessed both retirees 10 years before their retirement as being more or less equal.

Wessels says if you want to ensure that the volatility of investment markets and investment cycles will not undermine your future financial security, you will have to reject calculations based on average returns and in all probability you will need to:

* Save more before retirement;

* Retire when you are older; and/or

* Withdraw a lower pension in retirement.

He says even without the effects of the vagaries of market returns, many people reach retirement age without sufficient capital because of inadequate savings rates, starting saving too late, or withdrawing retirement capital when changing employers.

“In short, bad choices are made, both deliberately and naively.”

And then there are those people who have had their retirement savings stolen along the way or in retirement by the unscrupulous.

“The stark reality for many people nearing retirement is that they cannot actually afford to retire – in any case, not at the time they are supposed to retire,” Wessels says.

Added to this is the fact that most people will only realise close to retirement that their savings will not be sufficient to provide an inflation-adjusted pension for 20 years or more.

Wessels says what is already a high-risk situation is exacerbated by using incorrect assumptions about investment returns. “Unfortunately, human behaviour dictates that we often extrapolate recent investment return experiences very far into the future, even if we know the mean reversion phenomenon exists.

“Perhaps it’s a case of hoping or desiring a certain outcome rather than having realistic return expectations,” he says.

The reality is very different. Real (after-inflation) returns of three main asset classes of equities, bonds and cash over the past 110 years have shown significant variations in returns (see the graph, “South African asset classes: rolling 10-year real returns 1900-2010”, link at the end of this article).

Wessels says returns are cyclical for all asset classes, with periods of high real returns followed by periods of low real returns, and with the volatility of returns being the most pronounced with equities. This volatility will have an impact on your investment portfolio.

Using the constant return assumption in your planning means you may mistakenly take it for granted that:

* Before retirement. You will about double your retirement wealth in the last 10 years before your retirement. This also assumes that you will continue to contribute to your savings at a rate of 15 to 16 percent of your income, with the balance coming from investment growth.

Wessels says this last 10 years of your working career is crucial in determining whether your retirement plan is sustainable. It will determine your pension for the rest of your life if you purchase a guaranteed annuity. In the case of an illa, it will determine the percentage of your capital you will be able to withdraw as an income. Your pension withdrawal rate is one of the factors that determines how long your capital will last in retirement.

* In retirement. You will be able to maintain the same inflation-adjusted pension, both as a percentage of your remaining capital and as a rand amount, from an illa, with the same constant impact on your retirement capital.

Wessels says, in short, the higher the long-term real returns, the lower your required savings rate, or, in the case of an illa, the higher the safe withdrawal rate – but “you won’t know this beforehand!”

He says that it is virtually impossible to predict an outcome 10 years before retirement, or even five years before retirement. It is simply the luck of the draw.

Wessels says his research shows that with a typical retirement investment portfolio, with 50- to 60-percent exposure to equities, you can “only achieve relative certainty about final retirement values two to three years before retirement”.

He says that even though many people nearing retirement cannot actually afford to retire when they are supposed to retire, not everyone is in a position to keep toiling for as long as they wish.

You may even find that your scheduled retirement date actually comes sooner because of being forced into early retirement by an employer rationalisation programme, or because your skills have become redundant, or because of ill health.

And, as a final warning, he says his research is based on “known unknowns”, based on the variable investment returns over the past 110 years.

“There is no guarantee the future will not perhaps be worse than what we have experienced so far.”

The key ingredients to make your retirement a successful financial journey lie in “conservatism and prudence in your retirement projections”, Wessels says.


Mean reversion: A theory suggesting that prices and returns eventually move back towards a mean or average. This mean or average can be the historical average of the price or return or another relevant average such as the growth in the economy or the average return of an industry. This theory has led to many investing strategies involving the purchase or sale of stocks or other securities whose recent performance has greatly differed from their historical averages. However, a change in returns could be a sign that the company no longer has the same prospects it once did, in which case it is less likely that mean reversion will occur. Percent returns and prices are not the only measures seen as mean reverting; interest rates or even the price-earnings ratio of a company can be subject to this phenomenon. (Source:

Volatility: A statistical measure of the dispersion of returns for a given security or market index, on both the up and downside. Volatility can be measured by using either the standard deviation or the variance between returns. The higher the volatility, or the investment’s propensity to move up or down in value, the greater the presumed risk. The risk for retirement fund members is that at retirement there could be a significant drop in market values, with a consequent knock-on effect on their accumulated retirement savings.

Replacement ratio (also known as the net replacement ratio or NRR): The amount by which your pension will replace your final pensionable income. For example, if your final income is R400 000 and your pension at retirement is R300 000, you have a replacement rate of 75 percent.


In undertaking his research, financial adviser Daniel Wessels used a number of variable factors to calculate how different levels of retirement wealth accumulated by age 55 will affect your maximum retirement or safe retirement age in worst-case scenarios.

A. Your years in retirement

(Refer to table A, link at the end of this article)

Wessels says the length of your planned post-retirement period will have a major effect on the most appropriate age at which to retire.

He says it is easy to justify an early retirement age if you do not use realistic projections.

Wessels warns that an earlier-than-normal retirement age, of say 55, is only appropriate if you have accumulated retirement capital worth, say, at least 20 times your final pensionable salary.

However, he says you need to be realistic about how long you will live. Many retirement plans are based on someone aged 60 dying in their mid-80s. But this is an average and you could live to be 100.

So assuming that you, at the current age of 55, have retirement savings of 20 times your pensionable salary, you will be able to afford to retire at your next birthday if you live to be 80, based on worst-case investment scenarios. However, if you live to be 100, on the same assumptions you should not retire before 65 so that you can accumulate sufficient capital to ensure you do not run out of money before you reach that above-average ripe old age.

B. Your savings rate

(Refer to table B, link at the end of this article)

Wessels says what you save as a percentage of your income from age 55 will make the least difference, based on worst-case scenarios, of the various options you can consider to ensure a financially secure retirement.

For example, say currently, at age 55, you have capital of 20 times your annual pensionable salary and you can choose to save between five percent and 20 percent of your income. If you chose to save 20 percent of your salary, this would enable you to retire only one year earlier, at age 60, versus age 61 if you saved five percent of your salary.

C. Increasing your investment risk

(Refer to tables C1 and C2, link at the end of this article)

The risk exposure is based on the volatility of the asset class. For example, although equities have historically provided the best returns, they have done so in the most inconsistent manner year after year. So the higher the proportion of equities you have in your portfolio, the greater the chance of being the victim of a worst-case scenario.

Against equities, cash investments have historically provided the lowest volatility, but cash hardly ever provides above-inflation and after-tax returns.

Wessels says your choice of investment portfolio and the volatility risk you accept will make a significant difference to the age at which you will be able to retire assuming worst-case investment scenarios, depending on how much you have saved as a multiple of your income at age 55.

If, at age 55, you have saved 15 times your current pensionable salary and have an investment portfolio of zero exposure to equities going forward, your safe retirement age would be age 76. If, however, you have had a medium-level risk exposure of 50 percent invested in equities, you could retire 10 years earlier at age 66.

As another example, if you, at age 55, have saved five times your current pensionable salary and have a medium-level risk exposure of 50 percent invested in equities, your earliest retirement age would be 77, but if you have saved 20 times your salary you could retire 16 years earlier, at age 61.

Wessels says choosing between a medium-risk and a high-risk portfolio will, however, have a far smaller impact on your safe retirement age than the amount you have saved by age 55, based on worst-case investment scenarios.

D. Your safe retirement age based on replacement rate – pension as a percentage of final salary

(Refer to table D, link at the end of this article)

If you want to be sure of a pension that will replace your final salary by 75 percent or more, you need to be sure of having saved a lot of money before you turn 55. If not, your chances of achieving a replacement rate of 75 percent in the worst of investment times are extremely low, no matter what solutions you apply in the years between age 55 and retirement.

The research by Wessels shows that you will most likely have to accept a significant cut in income or keep working into your 70s.

Wessels says your final salary replacement rate is critical to your financial stability in retirement.

“Undoubtedly, if pitched too high, the chances are that your plan will not be sustainable for a prolonged period after retirement,” he says.

So, say at age 55, you have saved 20 times your annual salary and are prepared to take a 75-percent cut in your monthly income (a replacement rate of 25 percent), then you can retire today. But if you want a 100-percent replacement rate and want to be sure of a sustainable income for the rest of your life, based on worst-case scenarios, you must work for another 12 years and retire at the age of 67.

And the calculations again show that the less you have saved by age 55, the lower your replacement rate will be and the longer you will have to work.

Wessels says the problem is that the decision about when you will retire is, more often than not, out of your hands.

This means, for example, that a person retiring at age 65 with accumulated retirement wealth of 15 times salary at age 55 should not base an illa pension on replacing more than 70 percent of his or her final salary.

Likewise, someone retiring at age 60 with accumulated retirement wealth of 20 times salary at age 55 should not aim to replace much more than 70 percent of final salary.


If, taking account of the worst possible historical scenarios, you want to be sure of a financially secure retirement and do not want to have to drastically reduce your standard of living in retirement, you will probably have to save more and/or delay retiring.

This is the conclusion of research undertaken by Daniel Wessels of financial advice company Martin Eksteen Jordaan Wessels.

He says most people will probably need to save more and delay their retirement.

Wessels says that instead of using the “constant return assumption”, your financial adviser should do an assessment of your worst-case scenario at least 10 years before your currently scheduled retirement date so that you can decide on what solutions are possible.

This assessment should take account of the worst-

case scenarios over the past 110 years, the period for which records are available for the annual performance of asset classes.

To illustrate the challenge of getting it right, Wessels uses the example of a woman, Ms X, aged 55 who, 10 years from retirement, is planning for retirement using the constant return assumption (see table, link at the end of this article).

By assuming a constant five-percent real (after-inflation) return over the next 10 years, Ms X could expect to grow her retirement fund to R7.27 million in today’s terms, or nearly 15 times her final annual salary.

At retirement, she expects an annual pension from an investment-linked living annuity (illa) of R375 000 a year (in today’s values). This would equate to an initial pension with a withdrawal rate of five percent. Ms X’s withdrawal rate is also based on maintaining a pension adjusted for inflation for 25 years after retirement, with her capital providing a constant real return of five percent a year.

Wessels says her plan would seem to provide a sustainable income in retirement, but it is based on constant return assumptions that are “unrealistic and that make the projections unrealistic”.

When back-tested against historical return data, Ms X’s retirement plan portrayed a less rosy picture.

In back-testing, Wessels used return data from 1900 to 2010 and based his calculations on the return profile of a medium-risk portfolio invested 50 percent in equities, 30 percent in bonds and 20 percent in cash. He created hypothetical sets of 35-year portfolios (10-year pre-retirement and 25-year post-retirement), starting in 1900 and ending in 2010. For example, the first set started in 1900 and ended in 1934, the second started in 1901 and ended in 1935, and the last set started in 1976 and ended in 2010 – 77 data sets in all.

Wessels found that Ms X’s money ran out at age 80 in 20 percent of the 77 scenarios. This increased to 30 percent by age 85 and 40 percent by age 90.

To avoid any disaster, Ms X could elect to increase her savings or retirement fund contribution rate over the 10 pre-retirement years from 12 percent to whatever rate would be required to nullify the risk of running out of retirement capital before reaching the age of 90.

But an analysis shows that this option would be unsustainable, because, to ensure zero chance of plan failure, the minimum savings rate required would be as high as 80 percent of her gross salary.

Wessels says the difference between a five-percent and 20-percent contribution rate would make a difference of sustaining her retirement for only an additional three years.

An alternative option would be for Ms X, if possible, to extend her retirement date beyond the planned 10 years, and keep on working and contributing to her retirement savings for as long as possible.

If Ms X was to experience similar return conditions as at the onset of the 1920s, she could retire today at age 55, because she would have sufficient retirement capital. But this would not be so if she was to experience returns similar to what retirees would have experienced at the onset of the 1940s and 1950s. If Ms X was planning her retirement under the investment market conditions of the 1940s and 1950s, she would need to continue working for a maximum period of 18 years from today, extending her planned retirement age by eight years – from 65 to 73.

These calculations were done on Ms X living in retirement until the age of 90 (a retirement of 25 years). However, Wessels says that no one knows in advance how long they will live in retirement. When changing the post-retirement period from 15 years (age 80) to 35 years (age 100), the maximum safe retirement age would have increased from 69 years to 77 years.

In short, he says that for every five years added to the post-retirement period, Ms X would have to work an additional four years.


The government, with its proposed National Social Security Fund, is targeting a pension for members that will provide them with a monthly income of 40 percent of their final pay cheque. Unfortunately, a monthly pension of 40 percent of final salary may also be the fate of more optimistic retirement fund members if they are faced with a worst-case investment scenario.

Daniel Wessels of financial advice company Martin Eksteen Jordaan Wessels says that most members of retirement funds should consider this worst-case scenario in their calculations when retiring at age 65.

The replacement rate (see “Definitions”, above) of 40 percent of your final pensionable salary could be the worst-case scenario outcome despite your having targeted a replacement rate of 75 percent based on a constant return assumption of what markets have done on average over the longer term.

Wessels says he is not implying a 40-percent replacement rate is always the worst-case scenario for everyone.

“It obviously depends on how much capital you accumulate. Where an individual has accumulated retirement savings equal to eight times salary 10 years prior to retirement, 40 percent would be the safe retirement rate at retirement.”

Most occupational retirement funds and financial planners target a replacement rate of between 70 and 80 percent at retirement based on between 40 and 45 years of contributions of at least 10 cents of every rand you earn.

Wessels says that if you want greater certainty of income in retirement that will ensure you a guaranteed income lasting until the day you die, rather consider a pension bought from a life assurance company than an investment-linked living annuity (illa), where you must accept the risk of the vagaries of investment markets as well as the risk of living so long that you outlive your retirement capital. This is especially the case where your initial withdrawal (drawdown) rate is in the order of seven percent or more.

He says that another obstacle that lurks for the unwary is that what to many people may seem to be a large sum of money on retirement day may, in fact, not be sufficient to provide a sustainable income in retirement, particularly if you live longer than the average retiree.

He says people who have little choice about when they retire will be better served if they ask themselves what is the maximum amount they should withdraw as a percentage of their final pensionable income that will ensure they will receive a sustainable income for life. This calculation should again be based on a worst-case scenario of what investment markets could do rather than on a constant return assumption.

He says that a worst-case scenario thrown up by his research for a person with retirement savings at age 55 of eight times annual salary would be a replacement rate at retirement of 40 percent of final salary.

This means that a maximum replacement rate of 40 percent should be targeted in the worst-case scenario when retiring at age 65.

“Obviously, such replacement rates will pose some tricky challenges for retirees, but it is imperative to budget and spend conservatively, especially during the first years of retirement, to minimise the likelihood of running out of capital during later years,” Wessels says.