If we take a look at the annual returns of the South African equity market over each of the last 113 years we can compare those returns to what an investor holding cash would have received.

The average annual difference, or excess return, that accrued to an equity investor is a significant 8.4%, which is why investors correctly talk of investing in equities for the long term. However, focusing solely on the average excess return excludes a lot of valuable information that an investor would want to know.

Over the 113-year period, the actual annual excess return from equities over cash has come within 2 percentage points of the 8.4% average on only six occasions, or just 5.3% of the time. As an investor it is important to realise that you are unlikely to receive the average return in any given year. In addition, the excess return has been negative on 44 occasions, or 39% of the time. There are also periods where there have been a number of these negative excess return years in close proximity. We are reminded of the Howard Marks (Oaktree Capital Management) saying: 'You don't want to be the six-foot man who drowned walking through a river that was fivefoot deep on average.' If we take a look at the annual returns of the South African equity market over each of the last 113 years we can compare those returns to what an investor holding cash would have received. The average annual difference, or excess return, that accrued to an equity investor is a significant 8.4%, which is why investors correctly talk of investing in equities for the long term. However, focusing solely on the average excess return excludes a lot of valuable information that an investor would want to know. Over the 113-year period, the actual annual excess return from equities over cash has come within 2 percentage points of the 8.4% average on only six occasions, or just 5.3% of the time. As an investor it is important to realise that you are unlikely to receive the average return in any given year. In addition, the excess return has been negative on 44 occasions, or 39% of the time. There are also periods where there have been a number of these negative excess return years in close proximity. We are reminded of the Howard Marks (Oaktree Capital Management) saying: 'You don't want to be the six-foot man who drowned walking through a river that was fivefoot deep on average.'

Focusing on the potential for equities to underperform cash in any given year may seem like a foreign concept to many investors in JSE-listed companies. Over the last 10 years the FTSE/JSE All Share Index (ALSI) has only underperformed cash twice (and barely so in 2011), with the remainder of the years producing an exceptional excess return of 13.2%, on average. History suggests this good fortune cannot continue indefinitely.

Readers of our factsheets and commentaries will know that for some time now we have been cautioning that we believe the South African market is overvalued. The cash and hedges we have held as a consequence of this view have been a drag on the Fund's returns. However, the published return on cash (the prevailing interest rate) understates its current value, as it ignores the option that cash gives you to increase exposure to equities at more attractive valuations. We believe this option is attractively valued at current market prices.

While we do not manage the Balanced Fund to a specific net equity exposure number (see our March 2014 commentary), we have the ability to increase the Fund's net equity exposure materially if we find attractive opportunities. The net exposure to equities remained roughly stable over the last quarter as a result of the pleasing increase in price of some of the Fund's big holdings such as Sasol, British American Tobacco and SABMiller, which we continue to find attractive relative to rand cash.

We patiently wait for the opportunity to increase exposure to equities, while aiming to limit the downside should we encounter a period of negative excess returns.

Commentary contributed by Duncan Artus

Happily for the Allan Gray Balanced Fund, Sasol was the Fund's largest investment over the past two years. The company accounts for 6.6% of Fund assets and close on 13.8% of the domestic equity, a sizeable position. The obvious question whenever a holding has outperformed the market is: 'Surely you should be selling?' Let's look at the numbers.

The financial year to 30 June has just passed and the company should earn about R60 for the period, compared to the share price of R630 and the FTSE/JSE All Share Index, which trades on 18.4 times historic earnings. But one shouldn't use point-in-time earnings to value a company; the important criteria are the through-the-cycle earnings and how management decides to invest the earnings that are not returned to shareholders as dividends.

When assessing the through-the-cycle earnings of Sasol, the important drivers are the rand/US dollar exchange rate and the oil price. The rand averaged R10.40/US$ over the past year, a level that we think is about normal. When looking at the probability of either substantial strength or weakness from here, we think the probability of the rand weakening is more likely. A trickier subject is the oil price, which has remained remarkably stable at around US$110/bbl over the past three and a half years. We think US$110 is on the high side and use a lower price (US$90) when estimating Sasol's normal earnings. The outcome of this calculation is a normal earnings level of R52 per share. This means Sasol is trading on 12 times normal earnings, compared to the market, which trades on 18.4 times what we think are high earnings. It is clear to see why Sasol is our favourite stock, despite the recent price appreciation.

Sasol's recent capital allocation track record is mixed at best, so it is concerning that it is looking at some very large growth projects. Fortunately management seems to be bringing a new cost and execution discipline to the business; the potential for cost saving is substantial-about R3bn (R3.25 per share), according to management. The cost focus is on simplicity and eliminating waste, with a small example being the target reduction in legal entities from 250 to 50. The simplicity and clearer lines of responsibility should allow for better capital allocation decisions. Sasol has two major projects on the drawing board. The ethane cracker they plan to build in Louisiana should offer good returns, a fast pay back and relatively low execution risk. The investment decision on the very large US gas-to-liquids project is still some way off, however, we think this project is unlikely to go ahead as, in our opinion, there are better and less capital-intensive ways to monetise US shale gas.

Our asset allocation decisions in the Balanced Fund are the result of a bottom-up process that entails assessing the attractiveness of the individual assets in every asset class (e.g. shares, bonds, property, commodities and cash) and combining them in a diversified portfolio. In other words, we don't look at how the FTSE/JSE All Share Index (ALSI) stacks up against the All Bond Index, but at the risk-adjusted return of every individual share versus the most attractive opportunities in alternative asset classes. In an environment where the stock market offers fewer attractive opportunities, the weighting of other assets naturally increases.

The ALSI briefly broke through its all-time high of 48 000 on the last day of the quarter. With the average company now on a price/earnings (PE) multiple of 17x, the risk is to the downside: the long-term average PE multiple is just under 12x. While we continue to find attractive shares, those are fewer and farther between or at smaller discounts to intrinsic value. Consequently, the proportion of the Fund allocated to South African equities has gradually been diminishing.

Of every R100 invested in the Fund, a net R55 is currently exposed to equities. R44 of the R55 is invested in shares that are listed in South Africa. This number is low compared to the Fund's history: since inception in October 1999, the net domestic share exposure has averaged 54%. Historically, it has only been lower than today 15% of the time.

This is one indication that we are currently more worried about the risk of capital loss from SA-listed shares than at most times in our past. But this number in isolation fails to portray a full picture of the Fund's sensitivity to stock-market fluctuations. One measure of the risk of a share in comparison to the market as a whole is called 'beta' in financial jargon. A beta of one indicates that the share moves exactly in line with the market. A beta of less than one means the share goes up less when the market goes up but also falls less when the market falls, and the opposite is true for a share with a beta of greater than one (a negative beta share moves in the opposite direction to the market).

The domestic shares in the Fund have a beta of 0.79 in aggregate, measured over the past six years. This means that if historic trends prevail, the Fund's domestic shares are less impacted by what happens on the SA stock market than the average share. This is largely due to the stability provided by shares such as BAT, Reinet, Remgro and Nampak.

Beta is a historic measure and will change over time. But we are hopeful that the shares we have picked in the Fund will outperform if market valuations fall back to their historic averages. Coupled with a conservative allocation to the various asset classes, we believe the Fund is well positioned to weather an uncertain future.

The Fund's investments outside South Africa contributed significantly to its returns in 2013. The FTSE World Index returned 24.7% (in US dollars) for the calendar year. Successful stock picking boosted the dollar return to 43.2% in the Orbis Global Equity Fund. As the rand depreciated from R8.40 per dollar to R10.35 per dollar over the year, the Fund's foreign returns were even higher when measured in rands. Although the Fund's investment in African (excluding SA) securities is still relatively modest, these investments also performed strongly, with the Allan Gray Africa ex-SA Equity Fund returning 20% in US dollars.

In light of this strong performance, Fund investors may be asking whether or not we should now be reducing the Fund's foreign exposure. To some extent we are required to sell in order to comply with Regulation 28 of the Pension Funds Act, which prescribes that foreign exposure which exceeds 25% due to market value movements must be reduced back below 25% within 12 months. Over the year, the Fund repatriated R2.5 billion from its foreign investments. (By the way, this is an illustration of the important stabilising effect of South Africa's new retirement fund regulations).

However, we don't judge it appropriate to reduce the Fund's foreign exposure significantly below 25% at this time. We continue to find more attractive stock-picking opportunities offshore than in South Africa.

Despite the weakness of the rand over the last three years, we remain concerned that the local currency is vulnerable, as South Africa relies on the regular inflow of foreign capital to pay for our current account and fiscal deficits. The export response to the weaker rand has been disappointing so far - the mining sector has been plagued by labour unrest and policy uncertainty, and many South African manufacturing enterprises are still struggling to be globally competitive. The Fund's foreign investments also provide diversification, which is particularly important given the limited universe of listed South African securities.

We are pleased that the Fund has substantially outperformed CPI inflation over the last year, and that it has outperformed its peer benchmark in a strong year for equity markets despite holding a below-average equity exposure. Many sentiment indicators are exhibiting extreme optimism, and equity valuation multiples have expanded. This should make absolute returns harder to achieve in 2014. Should the markets sell off, we hope that Fund investors will be rewarded for the Fund's cautious positioning over the last three years by the Fund weathering the storm better than its peer benchmark.