The benchmark FTSE/JSE All Share Index (ALSI) declined by 6.9% in the third quarter, but this performance was flattered by the weakness in the rand. In US dollar terms, the index declined by 22.2% for the quarter. The resources sector declined more than the overall market. The dollar prices for base metals and the platinum group metals fell significantly, but they remain substantially above their long-term averages.

Despite a decade-long commodity bull market, Anglo American has underperformed the ALSI over the last 10 years. It is now trading not far away from its relative low over the last decade. The investment case for Anglo American is mixed. For example: we are concerned that the rapid growth in credit and the commodity-intensive building boom in China are unsustainable, but we are also concerned about the sustainability of the rand's strength. We believe that Anglo's base metal expansion projects in South America will yield attractive returns through the commodity cycle, but we are concerned that its investment in the Minas Rio iron ore mine and its associated infrastructure in Brazil will not deliver the returns the company expects.

A further example of the mixed investment case stems from the conglomerate nature of Anglos. One can think of buying one share in Anglo American for R275 (the quarter-end price) as the equivalent of investing R93 in Anglo American Platinum and R72 in Kumba Iron Ore (both separately listed on the JSE), and R110 in the remaining (or 'stub') assets of Anglo American. Anglo's investment in Kumba has proven to be a spectacular success, but at its current price we believe the risk is to the downside. This may seem surprising with Kumba trading on less than 10 times profits, but our concern is that Kumba's current profitability (72% EBITDA margins) is unsustainably high. Kumba currently generates higher operating profits per employee and contractor (US$417 000 pa) than does Goldman Sachs.

While the platinum miners have all underperformed the market significantly, we regard Impala Platinum as a more attractive long-term investment than Anglo American Platinum at current prices. Although a large proportion of Impala's profits is currently being spent on sinking new shafts, this should entrench the company's advantage as a low-cost producer over the longer term. Impala is now a top 10 holding.

On the other hand, we regard R110 as an attractive price to pay for Anglo's remaining 'stub' assets, which include De Beers, its copper and nickel mines in South America, its coal mines in Australia, South Africa and Colombia and other non-core assets. We have thus skewed the Fund's position in Anglo American towards the 'stub' assets by investing a portion of the total position in Anglos 'stub' certificates, which are listed on the JSE and priced daily.

There is widespread optimism on the continuing growth of the South African consumer. A minority share in Massmart is now trading close to the premium price Walmart recently paid for control of the dominant South African mass merchandiser (equivalent to 26 times earnings). But what about the South African producer? South Africans' consumption binge can continue for as long as creditors (especially foreign creditors) are willing to fund private sector credit extension and public sector deficits. But in the long run, consumption will need to be funded by production. Not all, but many of the companies which harness the efforts of South Africans to produce goods and services to trade with other countries are presently struggling with many challenges including rapidly rising electricity and labour costs, an inflexible labour regime and a strong rand. These challenges are well known so this is a natural place to look for value that the market is missing, and we have added to our positions in a few selected mining companies. If these challenges prove to harm the long-term productive potential of South African companies and their employees, then they will most probably harm their long-term potential to consume too. The Fund continues to be significantly underweight South African retail shares.

The sector in which the Fund is most underweight is the basic materials sector. Iron ore is currently a significant profit contributor for the diversified miners. We see little upside to iron ore prices from current levels, and see no reason why the current super-profits earned by iron ore miners will not be competed away.

The Fund's second most underweight position is in the consumer services sector. This is partly owing to the Fund having a negligible position in the retailers. In the early 2000s, the retailers accounted for some of the Fund's biggest overweight positions. At that time business conditions for the retailers had been poor for some time, and the low P/E multiples at which the shares traded suggest that the market believed that these poor conditions would persist. This belief proved unfounded as the retailers entered a fantastic boom in which their profits multiplied by factors ranging from four to 10 times.

Today it appears that the market expects these favourable conditions to persist, as the retailers are generally trading on high P/E multiples. But we find it hard to envisage how things can get much better for the retailers, which makes it hard to see much upside potential from the current share prices. On the other hand we perceive downside risks, which the market is choosing to ignore. These risks include considerations such as whether the prevailing trend of above-inflation wage increases is sustainable in an economy with widespread unemployment and constrained growth potential, higher inflation and interest rates, aggressive unsecured credit extension, and unusually high returns on capital for the retailers.

Since the year 2000 when it traded as low as US$6 per share, SABMiller has convincingly outperformed the MSCI World Index. Today SABMiller is trading around US$34.50 per share, while the MSCI World Index is still down on the average level at which it traded in the year 2000.

Over the last two years SABMiller has continued to beat the MSCI World Index by a wide margin. It traded below US$13 per share in March 2009, before bouncing back to a new dollar high in December 2010.

If we were holding SABMiller in a global equity fund, it would certainly be tempting to lock in some profits and lighten exposure at current prices. But this equity fund is a South African fund, not a global fund. And quite incredibly, SABMiller has performed no better than the benchmark FTSE/JSE All Share Index over the last two calendar years.

This speaks to just how strongly the South African stock market has performed of late. Since its lows in late 2008, it has almost tripled in US dollar terms. It is now back to its dollar highs of October 2007. Fund investors should not forget that after peaking in October 2007, this benchmark index lost two-thirds of its dollar value in just under a year.

SABMiller is trading on a relatively full multiple of 19-20 times adjusted earnings, but we believe these earnings will prove to be more sustainable in the long term than the current above-normal dollar profits being earned by iron ore and copper mines, and South African retailers. In addition, SABMiller has a number of opportunities to steadily grow its profits for many years by investing in its own business at high returns on equity - some examples include ongoing efficiency initiatives (especially in its North and Latin American operations); long-term growth in per capita consumption of beer in emerging markets (especially in Africa and India); and growing volumes, market share and pricing power in China.

So although SABMiller is not as attractively valued as we would like in absolute terms, we still consider it to be relatively attractive compared to most alternatives in our investment universe on the JSE, and the Fund is currently maintaining its full position in SABMiller.

Should investors wish to access a broader (and global) investment universe, they can do so through the Allan Gray-Orbis foreign unit trusts.