Choosing the right level of income annually represents a unique opportunity for investors to manage their risk of ruin.

Jun 28, 2018

5 minutes

Choosing the right level of income annually represents a unique opportunity for investors to manage their risk of ruin.

This article is the second in our series on how to manage a living annuity to provide an inflation-proof income over a period of 30 years.

In our first article we introduced some of the key conclusions from our living annuity research, and we expanded on how important volatility is to the long-term success of an income portfolio.

In this second article, we take a closer look at one of the other key components of the pensioners’ puzzle – how to manage the income withdrawal strategy for a living annuity investor.

From our research work we have established that a successful living annuity income strategy contains two key components:

- Picking a safe income level at the outset.
- Adjusting the income every year to reflect investment returns achieved on the annuity, whilst preserving the real purchasing power of the income.

Below we use the results from our in-house living annuity research model to expand on both these points. You can find more detail on our living annuity model here (investment return model, income strategies, fees, definition of a successful annuity etc.).

The level of inflation-growing income an investment portfolio can provide has been the subject of a fair amount of academic research over the past 25 years.

The work of Bengen (1994) on sustainable incomes from US equity portfolios represents the starting point for many subsequent research papers into this question. Bengen’s paper concluded that US investors could draw a 4% income from a US bond and equity portfolio for life, with little risk of ever running out of money.

To investigate the question from a South African pensioner’s perspective, we adapted our in-house living annuity model to:

- Calculate optimal investment portfolios for a range of living annuity starting incomes (i.e. the best combination of cash, bonds, equities and international assets that minimises the probability of the annuity not producing a 30-year inflation-proof income.)
- Measure the annuity’s ability to produce an inflation-proof income under different initial income scenarios, as well as different annual income review strategies.

Our starting point income scenario was the baseline strategy: the pensioner chooses a starting income, and then increases their income from the annuity by exactly inflation every year.

Figure 1 illustrates the failure rates of such an income strategy. From the graph we can see the following:

- Initial incomes up to 4% of starting capital result in virtually no failures over the 118 years that we simulated in the model.
- Annuities start failing from above a 4% starting income, and at a 5% initial income, the failure rate approaches 10% of annuities.
- Starting incomes larger than 5% experience rapidly increasing failure rates, and at around a 7% income, we see failure rates in excess of 60%.

**Figure 1: Income strategy failure rates**

Source: Ninety One SA (Pty) Ltd.

The conclusion from this analysis supports the results from international academic papers on sustainable incomes that a 4% initial income level represents the limit of a ‘safe withdrawal’.

But what about other options? Can a pensioner improve their odds if they approach their income increases in a different way? We proceeded to investigate these questions.

When we evaluated different income review strategies in our model, it became apparent that a pensioner can substantially improve on the fixed inflation-adjusted income scenario. Other than reducing the initial income selected, the simplest way to improve the sustainability of a living annuity is to link the annual income increase to the level of investment performance achieved during the preceding year.

A simple way to link the annual income increase to the portfolio return is the following algorithm: Increase the income every year by the net capital growth in the living annuity portfolio (after income and fees), but keep the annual income increase between inflation minus 5% and inflation plus 5%. Capping the increase above and below inflation smooths the annual increases and is more reflective of the reality that very few pensioners can afford a substantial cut in their income.

If we model this amended income review strategy relative to our first strategy, the results are as illustrated in Figure 2.

**Figure 2: The impact of different income withdrawal strategies **

Source: Ninety One SA (Pty) Ltd.

Figure 2 shows how our amended income review strategy substantially improves the sustainability of the living annuity, virtually halving the probability of failure in the critical 5% to 6.5% initial income range.

Our original income strategy produced failed annuities as starting incomes moved above 4% p.a. In contrast, our revised income review strategy only starts failing at starting income levels exceeding 5% p.a. The amended strategy delivers a 20% increase in initial income in rand terms, but still produces a fairly ‘safe’ annuity.

Pensioners and advisors will acknowledge, however, that it is difficult to plan retirement expenses when annual pension increases can fluctuate wildly. It is for this reason that we capped both the minimum and maximum annual increases in our example. A financial advisor can play an important role by guiding pensioners with their pension increases every year.

A pensioner’s income withdrawal strategy is one of the key drivers of the sustainability of a living annuity. A good income strategy achieves the following objectives:

- The pensioner chooses a sustainable starting income level (i.e. 5% of capital or less)
- The pensioner links their annual increase to the performance achieved on the annuity’s investment portfolio, but the income broadly keeps pace with inflation.

In the next instalment in our series, we will investigate the other key driver of a living annuity’s success – the construction of the annuity’s investment portfolio.