With charges in the retirement industry under scrutiny, and criticism against prescribed asset allocations in some circles, investors may be wondering if they should rather construct their own investment portfolio for retirement.

While a retirement annuity (RA) offers tax benefits, some argue that the potential out-performance of a discretionary investment that does not have to adhere to prescribed asset allocations (Regulation 28), such as a maximum 75% allocation to equities, would offset any tax benefits gained from using an investment vehicle like an RA.

Daniel Wessels, a financial advisor at Martin Eksteen Jordaan Wessels in Cape Town, has crunched some numbers. While the findings make for some thought- provoking reading, the numbers should be looked at with caution. There are various factors the model does not take into account, such as the potential impact of a lack of investment discipline, or capital gains tax and estate duty on a discretionary investment.

The model

The analysis tried to determine what the out-performance of a discretionary investment portfolio have to be over a certain time period to compensate for the tax deductions that an RA investor would qualify for. In other words, the after-tax income in retirement of both portfolios had to be the same. Therefore, when considering a discretionary investment portfolio instead of an RA investment, one should bear in mind that it must outperform an RA investment on an after-tax basis.

The model assumes that a maximum of 15% of the RA investor’s contributions are tax-deductible. This concession also needs to be balanced against the fact that, in retirement, the annuity income of such a portfolio would be 100% taxable at the individual’s marginal rate. In the case of a discretionary investment, the same contributions would be made (i.e. 15%), less the tax savings forgone by not making use of the allowed deductible contributions in an RA investment.

The tax aspect gets a bit complicated, however, in that whereas the growth in a RA investment portfolio bears no tax implications, the discretionary investment returns are subject to taxable interest, dividend tax and capital gains tax.

At retirement, the discretionary portfolio would be restructured to provide for an income stream to the retiree, and it is assumed that between 25% and 50% (depending on the split between interest, dividends, and capital-and the tax implications thereof) of the income would be taxable at the marginal tax rate.

Wessels says that while some may argue that a portfolio with a 100% equity allocation would deliver much better results than a portfolio with a 75% equity allocation, a longer term history does not provide convincing evidence of that being the case, adding that the recent stellar returns in the local equity market, have skewed perceptions in favour of the equity market. However, if one analyses data since 1926, there are also long periods of time where the equity market did not perform that well. It is therefore important to take a long-term view.

In summary, the model tries to determine what out-performance the discretionary investment will require in order to result in the same real after-tax return in retirement.

Crunching the numbers

The model suggests that an RA is most beneficial to investors in the medium to high income categories. Stated differently, the discretionary portfolio would need to outperform the RA by up to 1% per annum to achieve the same income result.

The RA is perhaps less beneficial to lower-income earners, because they are typically in the lowest marginal income tax bracket of 18 per cent. A lower-income individual whose annual contributions to the discretionary investment amount to R15 000 therefore only has to achieve a marginally higher annual return (between 0.2 and 0.4%) than their RA counterpart to get the same after-tax income in retirement.

If the same variables apply, but the rate of 15% is increased to 27.5% (the rate that will apply from an effective date, on or after 2015), the findings are the same, Wessels says. In other words, the amount contributed does not impact the results provided that the same is contributed in both cases.

However, if the model is adapted to assume that the discretionary investment portfolio consistently out-performs the RA by 1% per annum (after tax), the findings also indicate that an RA is more beneficial to mid- to high-income earners, but in that scenario the discretionary investment would be better for both low-income and high net worth individuals .

A word of caution

It is important to distinguish between new-generation RA products and traditional RA investments-the latter type of product is typically expensive and inflexible.

There are also other factors to take into account when deciding which route to go. Irrespective of the outcome of quantitative comparisons between an RA and discretionary investment options, perhaps the most important, non-quantifiable aspect to consider is that the discipline of a retirement annuity structure will invariably protect yourself against your own fallible investment behaviour, as you will experience the rollercoaster of investment emotions over time, by far the majority of people will benefit greatly by saving for retirement by using an approved retirement funding structure such as a retirement annuity product. In only exceptional circumstances, people with considerable wealth may be better off after-tax by structuring their own discretionary investment portfolio, and additional factors such as potential estate duties must be carefully considered.