There is this strange phenomenon in SA called Regulation 28 that is applied to retirement funds. It stipulates that retirement fund members may not have more than 75% of their funds in equities and no more than 25% of the fund invested offshoreIronically, Reg 28 is enforced at individual investor level in the accumulation phase (pre-retirement) and not to the draw-down phase (post retirement). Even so, ASISA has issued “guidelines” for “safe” asset allocation on living annuities. Not surprisingly, they have merely overlaid the Reg 28 allocations.
Historically, we have tended to err on the “cautious” side when it comes to Living Annuities; clients are dependent on the capital for their income. In this scenario, we have heeded the ASISA recommendations and have tended to keep offshore exposure at ±25-30% . The grounds for this are that the offshore portion of the fund is exposed to an additional risk on top of market risk, namely currency risk. The living annuity is therefore subject to a potential double negative effect of a falling market and strengthening currency (it has happened before)!
Recently, while looking at a client’s living annuity where they had made a decision to invest the entire amount offshore and to draw a 10% income from the fund (there is a long reason for this) I was very surprised to find that despite drawing 10%, the capital value had increased significantly since inception.
I know that the MSCI World Index has hit record highs recently but the rand is stronger (not weaker) than it was 1 and 2 years ago. During the same time the JSE All Share Index has been pretty much flat…most of the SA market has failed to give positive, never mind real returns, of any form.
This got me thinking about the ASISA guidelines and our “traditional and safe” approach to the allocation of capital within a living annuity. Perhaps it is not so safe after all?
Consider the following case for investing a greater portion offshore:
- The SA market is very small and very narrow – there are ±130 shares in the All Share Index as measured by Satrix. The MSCI All Country World Index consists of almost 1400 shares.
- If we look at the whole of the JSE, then the top 15 shares make up 63% of the entire SA market. In the MSCI World Index, the top 15 shares make up ±14% of the index. (The top 5 shares in SA make up ±50% of the Top 40 index).
- The largest SA share, Naspers, makes up ±19% of the ASLI while the largest share on the MSCI (Apple) is less than 2% of the index (its market cap is also greater than the entire JSE market cap).
- The whole of the SA market is less than 1% of the MSCI All country world index and yet most South Africans have around 90% or more of their total assets invested locally.
- On top of this we also have inflation risk which means that over the long term, the trend for the ZAR will be weaker (although, as history has shown, the decline is neither predictable nor linear).
I always tell clients to imagine themselves sitting on the moon and looking at earth as their investment option – would they put 99% of their assets into a market that is less than 1% of the total available market?
When considered like this, then there has to be a massive diversification benefit to investing offshore. There must surely be a very good case for investing more than 25% of a living annuity offshore too, especially if the annuitant is drawing a low income and can afford more volatility risk? (Available evidence seems to suggest that even those who are drawing a higher income could benefit from a greater offshore weighting too.)
And this is all before we even consider the political risk in SA, which has increased significantly and suggests that South African investments are very vulnerable going forward.
So what about it, are there any rocket scientists out there who can add anything meaningful to this? Is there a good case to be made for investing more than the “traditional” 25% of a living annuity offshore?
Notes: Thanks to Adrian Clayton (Northstar Asset Management) and Cassie Treurnicht (Gryphon Asset Management) for their help with the data.