GLOBAL ECONOMICS FOCUS

GLOBAL ECONOMICS FOCUS Commodity investors are being misled by historic returns 4 th Sept. 6 The historical returns on commodity futures appear attractive. However, in this Focus we look at the factors behind these returns, and find strong reasons why they cannot be expected to continue. One statistic often cited by commodity bulls is that even before the start of the current boom, investing in commodity futures gave a real return very similar to that on US equities. But almost the whole of this return was because futures prices were generally below spot prices (backwardation). Prices of the commodities themselves declined in real terms. This backwardation has now disappeared, so the prospective return on commodity futures will now be more similar to the return on the commodities themselves. This means that the appropriate historical comparison is with the negative real returns made on the underlying commodities. Indeed, many futures are now priced above spot levels (contango). This introduces an important headwind for investors, since underlying spot commodity prices now need to rise significantly if investment in futures is even to break even. There would be good reasons to expect a backwardation whilst speculation in commodity markets was only a niche interest. Correspondingly, it should come as no surprise that the backwardation has disappeared as mainstream investors started moving into commodities. The flood of investors into commodities appears already to have destroyed the main source of the earlier returns. Bulls argue that commodity prices will carry on rising, with supply struggling to catch up with increasing demand (in particular from China). But commodity prices are already unsustainably high, boosted by the Chinese investment boom and increased speculative demand. Investors thus face a double whammy, with the costs of the contango compounded by a sharp fall in spot prices. Diversification benefits are also likely to be weaker than in the past. Previously, inflation tended to boost commodities whilst undermining other assets. But inflation is not the risk that it used to be. Instead, the key risks over coming years are (a) slower global growth, and (b) reduced investor risk appetite. These factors are likely to undermine commodities at least as much as other assets. Capital Economics Ltd., 15 Buckingham Palace Road, London, SW1W 9TR. Tel +44 7823 5 Fax +44 7823 6666 www.capitaleconomics.com Managing Director: Roger Bootle (roger.bootle@capitaleconomics.com) Chief International Economist: Julian Jessop (julian.jessop@capitaleconomics.com) Chief European Economist: Jonathan Loynes (jonathan.loynes@capitaleconomics.com) Senior International Economist: Simon Hayley (simon.hayley@capitaleconomics.com) Senior US Economist: Paul Ashworth (paul.ashworth@capitaleconomics.com) International Economist: Keith Gyles (keith.gyles@capitaleconomics.com) Simon Hayley Global Economics Focus 1

Commodity investors are being misled by historic returns With equity returns continuing to disappoint, the historical returns on commodity futures look attractive. As a result, increasing numbers of investors are now looking at commodities as an additional asset class for the long term. However, in this Focus we look at the factors behind these returns, and find strong reasons why they cannot be expected to continue. We have already explored the outlook for commodity prices in our earlier Global Economics Focus Disappointing Chinese demand will further weaken commodities (sent to clients on 21 st June). We concluded that demand for key base metals in particular had been boosted by the unsustainable Chinese investment boom, coupled with speculative demand. As these forces wane, we expect prices to fall sharply over the coming year. This remains our view, and we do not re-visit these issues here. Moreover, we examined long-term price trends in another Global Economics Focus ( Commodity prices: current highs are not sustainable, sent to clients on 5 th May 6). We found that on average, real commodity prices had been on a sustained downtrend. (See Chart 1.) CHART 1: CRB COMMODITY PRICES (REAL, JAN. 1947=1) 1 CRB Spot Index Agriculturals Metals 1 1947 1957 1967 1977 1987 1997 7 Source CRB, Capital Economics. Deflated by US CPI. This immediately raises the question: how could investors have made attractive returns if commodity prices were falling in real terms? We address this question in the current Focus, and assess the outlook for future returns. To answer this question we turn to one of the most comprehensive surveys of the investment returns in commodity futures by Gary Gorton and Geert Rouwenhorst 1. This paper is often cited by those who recommend such investments, but when we look at the underlying data, it is less than encouraging. The key results are set out in Table 1. TABLE 1: HISTORICAL RETURNS TO COMMODITY INVESTING (JULY 1959-DECEMBER 4) Average annualised return (%) Commodity Futures 1.31 Spot Commodities 3.47 Inflation (US CPI) 4.13 Source Gorton/Rouwenhorst. Cumulative returns on up to 36 different commodities. Commodity futures did indeed give a very attractive 1.3% return. This was very similar to that achieved by investing in the S&P5 equities index. Furthermore, this excludes the huge returns that would have been made since December 4 as commodity prices boomed. But this data also shows that spot commodity prices rose by slightly less than the rate of inflation. This confirms that there was indeed a sustained downtrend in real commodity prices. The mystery is that commodity futures gave a much higher return than investment in the 1 Facts and Fantasies about Commodity Futures by Gary Gorton, University of Pennsylvania and K. Geert Rouwenhorst, Yale School of Management, June 4, updated February 5. Global Economics Focus 2

underlying commodities themselves. To explain this we need to look at the mechanics of investing in futures. CHART 2: COPPER FUTURES (CENTS/POUND, AUGUST 5) 1 17 15 13 Sep-5 Nov-5 Jan-6 Mar-6 May-6 Jul-6 Sources Comex, Bloomberg Chart 2 shows the prices of copper futures quoted on the COMEX exchange in August 5. Longerdated futures contracts were trading at prices well below short-term futures and the spot price. An investor buying a future for August 6 at 145 cents/lb would have seen its value shoot up to 344 cents/lb by the time it matured the following year. There are two components to this return. The largest is that spot copper prices rose sharply over this period (from 173 to 344). But the investor would have made a good return even if spot copper prices had stayed unchanged, because the future was bought at well below the spot price (giving a return of 19% in this case: 173/145-1). Such a downward slope in the futures market is known as a backwardation. Over the forty-five years covered by the Gorton/Rouwenhorst study, commodity futures markets have on average tended to be in backwardation. Futures contracts have only limited lives, so investors must reinvest in a new contract whenever an old one expires. The backwardation implied a cumulative roll yield, and over the years a substantial cumulative return has accrued, since investors would have tended to gain from the backwardation each time they rolled their investment on. Almost the whole of the historical 1 17 15 13 real return to investing in commodity futures has been due to this backwardation. This is the main factor explaining why investing in futures gave strong returns even whilst the commodities themselves declined in real terms. But, as we shall see below, this backwardation has now disappeared. Another, more minor, factor stems from the fact that futures can be bought on margin, thus obtaining a large exposure with only a limited outlay. In calculating these returns it has been assumed that the spare cash was invested in US government bonds. This was necessary in order to make these calculated returns comparable to those achieved on other assets, which cannot be bought on margin. But the 197s and 19s saw exceptionally high bond yields, and these boosted this average yield to around 7% over this period as a whole. With US yields now back below 5%, this is another component of the historical returns that we know will not be available to those who are investing in commodity futures now. Thus the key factors which were behind the attractive historic returns on commodity futures have both disappeared. One does not need to share our bearish view on the outlook for commodity prices even on a neutral view investing in commodity futures can no longer be expected to generate positive returns. Trends in the futures markets In looking at the degree of backwardation, we are clearly making a generalisation about the futures prices of many different commodities. One way round this is to look at the futures that are quoted on popular commodity indices. The most obvious example is the CRB commodity futures index, which covers 17 different commodities. (The compositions of this and other indices are shown in Table 2.) Comparing the spot price of this index with the price of the futures contract five months further out Global Economics Focus 3

gives us one measure of the backwardation (we chose the longest period that allowed a suitably long track record). This measure has fluctuated substantially over this period, but since the end of 5 it has moved to a substantial contango. (See Chart 3.) 1.8 1.6 1.4 1.2 1..98.96.94 CHART 3: CRB FUTURES BACKWARDATION/CONTANGO 1988 1991 1994 1997 3 6 Source CRB Contango Backwardation 1.8 1.6 1.4 1.2 1..98.96.94 This will have a substantial impact on investors. With futures five months ahead currently just less than 2% more expensive than the spot price, if this spot price were to stay unchanged, then investors would face an annual cost of almost 5% from investing in these futures. Spot prices would have to increase more rapidly than this for the investor even to break even. The impact of the backwardation can be seen in the annual total return (calculated to include the effect of rolling the contracts) compared to the annual percentage change in the index itself (which excludes rolling). Over significant periods in the past the total return was substantially boosted by the effect of the backwardation. (See Chart 4.) By contrast, the recent contango means that futures returns have been pushed almost 1% lower. As underlying spot prices have jumped by over % over the past year, the total return has still been excellent, but the negative return on rolling futures contracts represents a headwind for investors: underlying commodity prices must rise by more than this for investors even to break even. CHART 4: CRB FUTURES INDEX & TOTAL RETURN (% Y/Y) 5 3 1-1 - -3 CRB Futures total return CRB Futures index 1984 1988 1992 1996 4 Source CRB Weights as at 3 August 6 We might be concerned that this result only applies to the particular batch of commodities in the CRB futures index. But cross-checking against the GSCI index shows that this is not the case. TABLE 2: COMPOSITION OF COMMODITY INDICES (%) Goldman Sachs CRB futures index (GSCI) METALS 12.4 23.5 Copper 4.4 5.9 Aluminium 3. - Gold 2. 5.9 Nickel 1.3 - Zinc 1.2 - Lead.3 - Silver.3 5.9 Platinum - 5.9 FOODS 14. 52.9 Wheat 3.6 5.9 Cattle 3.2 5.9 Corn 2.2 5.9 Hogs 1.5 5.9 Sugar 1.3 5.9 Soy beans 1.3 5.9 Coffee.7 5.9 Cocoa.2 5.9 Orange Juice - 5.9 NON-FOOD AG..9 5.9 Cotton.9 5.9 ENERGY 72.8 17.6 Crude oil 46. 5.9 Other fuels 12. - Natural gas 6.4 5.9 Heating oil 8.4 5.9 Weights as at 1 September 6 5 3 1-1 - -3 The composition of the GSCI index is very different, in particular since it contains a high weighting on energy. (See Table 2.) But the futures Global Economics Focus 4

on this index have also recently moved to a large contango, especially as oil prices have fallen back over recent weeks. (See Chart 5.) these futures has already become negative. (See Chart 7.) CHART 7: GSCI TOTAL RETURN VS. SPOT RETURN (% Y/Y) CHART 5: GSCI BACKWARDATION/CONTANGO 1.1 1.1 1.5 Contango 1.5 1. 1..95 Backwardation.95.9.9 1992 1994 1996 1998 2 4 6 Spot - Total Return - Jan 1 Jan 2 Jan 3 Jan 4 Jan 5 Jan 6 Sources Goldman Sachs, Bloomberg - - Sources Goldman Sachs, Bloomberg. The impact can be seen in the GSCI return indices. The total return index includes the cost of rolling all the futures in the basket just before expiry. The backwardation on the futures contracts used to mean that this was a much more profitable strategy than simply holding the same commodities outright (spot). But this has ceased to be the case as the futures moved into contango. (See Chart 6.) Futures returns have been more than 1% below spot returns over the past year, and this differential is set to increase as the full impact of the rising contango feeds into the annual returns. The current even higher contango would imply a far higher annual differential. CHART 6: GSCI TOTAL RETURN LESS SPOT RETURN (% Y/Y) % 3% % 1% % -1% % 3% % 1% -1% -% -% 197 1975 19 1985 199 1995 5 Sources Goldman Sachs, Bloomberg. Annual returns on both futures and spot have fallen back as commodity prices have ended their previous spectacular rise, but the extra cost of the contango means that the total return to holding % Explaining the backwardation Thus far we have taken the degree of backwardation or contango as a given, and merely calculated the effects that this will have on investment returns. But it is worth exploring why futures markets tend to be in backwardation or contango. At first sight it may seem odd that there should ever be a persistent differential between future and spot prices. But there is a convincing explanation for the previous presence of a backwardation and its recent disappearance. There are two different explanations for the backwardation. Some argue that it encapsulates market expectations for the future. On this explanation, the habitual backwardation implies that market participants have, on average, generally been bearish. But the shift to the current contango would mean that investors entering the market are now having to buy in at high prices that not only reflect the huge run-up in spot prices over the past two years, but also the further expected rise that is priced into the futures market. We do not regard this as a good explanation of the backwardation. Instead, like other markets, increased bullishness should to a large extent be expected to be reflected immediately in the spot price, rather than in the differential between futures and spot prices. Global Economics Focus 5

One major item affecting the costs of holding physical stocks rather than futures is the financing and storage costs associated with holding a physical stock of commodities. Indeed, in the futures markets for financial products, this relative carry cost is the only factor which drives the spot/futures differential, and any deviations from this rule would rapidly be arbitraged away (eg. exchange rate futures simply reflect the interest rate differential between the two currencies). But for physical commodities the process of arbitraging between spot and future prices is rather messier, so this relationship does not hold so well. After all, with storage costs pretty constant, this could not explain the recent shift from backwardation to contango. Instead, the better explanation lies in the theory of normal backwardation, originally suggested by Keynes in the 193s. This focuses on the relative demand for hedging in the futures market. Commodity producers have a huge need to hedge by selling their output in advance in the futures market. This helps protect them against price fluctuations. But consumers of commodities can afford to be more relaxed, since any particular commodity is likely to be only one element of their costs (whereas it typically accounts for the whole of a producer s revenue). This means that there would naturally be an excess of players trying to be short in the futures market. This would lead to a persistent backwardation. Naturally, speculators might be attracted into taking long positions in these markets. Unlike the hedgers, speculators would be taking additional risk by participating in this market. They will generally only be willing to do this if they can expect a positive return from this, so even after the entry of speculators, commodity futures are likely over time to continue to show some backwardation on average. This normal backwardation is due to the relative scarcity of speculators willing to take long futures positions. But with mainstream investors becoming increasingly keen to include commodities in their strategic asset allocation, the previous shortage of natural long investors has been removed. Thus it should come as no surprise that the previous backwardation has now disappeared. This is an important point. When speculation in commodity markets was only a niche interest, these speculators were in relatively short supply, and earned an attractive average return. But the flood of additional money into commodity futures has removed this shortage and the associated backwardation. As we saw above, this backwardation was the source of almost the whole of the historic returns made by commodity futures. The flood of investors into commodities appears already to have destroyed the main source of the earlier returns. Diversification The other major attraction that is claimed for commodities is that the returns have tended to show a very low (and sometimes negative) correlation with the returns on other asset classes. Thus investment in commodities is expected to bring significant diversification benefits. But when we look into the reasons for this we find that the outlook is less encouraging. CHART 8: CORRELATION OF EQUITY & COMMODITY RETURNS 1..8.6.4.2. -.2 -.4 -.6 1961 1965 1969 1973 1977 1981 1985 1989 1993 1997 1 5 Sources Bloomberg, CRB. Rolling two year correlation of monthly returns on CRB futures index and S&P5. Over the past 45 years, the correlation with US equity returns has indeed been very low on average. (See Chart 8.) Indeed, the correlation was negative when general inflation was picking up in the late 197s and around 199. This fits with the 1..8.6.4.2. -.2 -.4 -.6 Global Economics Focus 6

idea that inflation tends to boost commodities whilst undermining other assets. But diversification benefits would not justify investing in any asset which we expect to give massively negative returns. Moreover, inflation is not remotely such a risk as it used to be. After all, we have just seen oil prices rise as much as in previous oil shocks (with crude rising from $25 per barrel in 3 to over $7 this year), yet this has not triggered the increased wage inflation that was seen following previous oil shocks. Nor is there any sign that inflation is expected to rise significantly: bond yields remain below 5% in all G7 countries. Instead, the key risks over coming years are (a) slower global growth, and (b) reduced investor risk appetite. These factors would tend to undermine commodities at least as much as other assets. This suggests that the diversification benefits of investing in commodities are also likely to be much weaker than in the past. Global Economics Focus 7