by Ke Tang and Wei Xiong, VoxEU.org
30 November 2010
In recent years, hundreds of billions of dollars of investment has flowed into commodities markets. This column describes why and how commodities markets have grown so rapidly and discusses some policy implications.
Crude oil, copper, cotton, soybeans, and live cattle â a seemingly unrelated set of commodities â went through a synchronised boom and bust cycle between 2006 and 2008 (see Figure 1).
Figure 1. Commodity prices
This cycle has stimulated heated debate in policy circles about whether speculation caused unwarranted increases in the cost of energy and food and induced excessive price volatility. There are two opposing views.
- One view puts the boom-and-bust cycle down to a simple matter of supply and demand.
According to Krugman (2008), Hamilton (2009), Kilian (2009), and others, the rapid growth of emerging economies such as China propelled the quick increase in world demand and caused commodity prices to soar before the summer of 2008. Prices later fell sharply as the world recession caused demand to fade.
- The second view points to the large flow of investment into commodity indices.
According to a CFTC staff report (2008) and Masters (2008), the total value of various commodity index-related instruments purchased by institutional investors has increased from an estimated $15 billion in 2003 to at least $200 billion in mid-2008. A recent report by the US Senate Permanent Subcommittee on Investigations (2009) argues that the dramatic index investment flow had distorted prices of some commodities such as wheat.
The growth of commodities
To thoroughly assess the role of financial investors in commodities markets, it is important to recognise the economic transition of commodities markets precipitated by the rapid growth of commodity index investment. We examine this process in a recent working paper (Tang and Xiong 2010). There is ample evidence suggesting that commodities markets were partially segmented from outside markets prior to early 2000s. Bessembinder (1992) and de Roon et al. (2000) show that commodity prices provided risk premium for idiosyncratic commodity price risk; Gorton and Rouwenhorst (2006) find that commodities had little price co-movements with stocks; and Erb and Harvey (2006) show that commodities in different sectors had little price correlation with each other. By arguing that commodities offer a diversification benefit to portfolios of stocks and bonds, Goldman Sachs and other indexers managed to promote commodity futures as a new asset class for institutional investors in the early 2000s following the collapse of the equity market. As a result, billions of dollars of investment has gradually flowed into commodities markets.
To identify the impact of increasing commodity index investment, we focus on the return correlations of non-energy commodities with crude oil. As index investors typically focus on their strategic asset allocations across different asset classes such as stocks, bonds, and commodities, they tend to move in and out of all commodities in a chosen commodity index at the same time. Their trading can thus cause prices of commodities in the index to move together. This effect is observed among stocks in the S&P 500 index (Barberis et al. 2005), and motivates our key hypothesis:
- Due to the increasing presence of commodity index investors after 2004, there is an increasing trend in the return correlations of non-energy commodities with crude oil, and the increasing trend is more pronounced for commodities in the two popular commodity indices: the S&P-Goldman Sachs Commodity Index (GSCI) and Dow Jones-UBS Commodity Index (DJ-UBS).
Figure 2 depicts the one-year rolling return correlations of crude oil with soybeans, cotton, live cattle, and copper. The correlations are computed from the daily returns of holding the front-month futures contracts of these commodities. The return correlations display a common feature in that they were all close to zero before early 2000s, but steadily climbed up after 2004 to above 0.5 in late 2009. Figure 3 depicts the average one-year rolling correlations among the 19 indexed commodities and the nine off-index commodities that have futures contracts traded in the US; indexed commodities are the ones which are listed in at least one of the GSCI and DJ-UBS indices. The two average correlations stayed at a stable level below 0.1 throughout the 1990s and early 2000s and were indistinguishable from each other. The average correlation among the indexed commodities climbed up to an unprecedented 0.5 in 2009, while the off-index commodities have only mild increases in correlation. This sharp contrast supports our hypothesis.
Figure 2. Rolling-return correlation of oil with different non-energy commodities
Figure 3. Average correlations of indexed and off-index commodities
The rapid economic growth of emerging economies such as China and India propelled growing demands for many commodities. To assess this effect, Figure 4 depicts front-month futures prices of 6 commodities â heating oil, copper, soybeans, wheat, corn, and cotton â in China and the US. Prices of heating oil, copper, and soybeans in China had boom-and-bust cycles closely matched with the corresponding cycles in the US. However, prices of wheat, corn, and cotton in China did not display any pronounced cycle, which is in sharp contrast to the cycles in the US. This contrast suggests that demands from China did not contribute to the price boom of all commodities.
Figure 4. Commodity prices in China and the US
The driver of the increased commodity return correlations in the recent years cannot be inflation because both inflation and inflation volatility remained low and stable since mid 1990s (see Figure 5). Neither could the driver be the financial crisis of 2008 as the increase in correlation and the difference-in-difference effect between the indexed and off-index commodities were significant even before the disruption of a full-scale financial crisis in September 2009. Furthermore, these effects were significant in different commodity sectors, such as grains, softs, livestock, and metals, and thus cannot be driven by growing production of biofuel in recent years, whose effect is limited to price increases of corn and its close substitutes in the grain sector.
Figure 5. Inflation and inflation volatility
Economic effects of financialisation
The increasing presence of index investors reduced the segmentation of commodities markets and can cause two types of effects.
- First, their presence could improve sharing of commodity price risk.
- Second, portfolio rebalancing of index investors can act as a channel to spill over shocks from outside to commodities markets and across different commodities, e.g., Kyle and Xiong (2001).
The data sample after 2004 is too short to give a reliable measure of changes in commodity risk premia, but sufficient for analysing the volatility spillover effect.
Figure 6 depicts the annualised daily return volatility of crude oil, the GSCI non-energy commodity return index, and the Morgan Stanley world equity index from 1989 to 2009. The figure shows that return volatility of non-energy commodities had been very stable at a level around 10% throughout 1990s and early 2000s. However, it started to rise after 2004 and peaked at an unprecedented level of 27% in 2008, which coincided with the hikes in volatility of oil and the world equity index. By decomposing daily returns of individual commodities, we find that the increase of their volatility was partially driven by their increased return correlations with oil; in particular, the increase in volatility was significantly greater for the indexed commodities.
Figure 6. Volatility of oil, GSCI non-energy index, and Morgan Stanley World Equity Index
Policy implications
In the aftermath of the synchronised boom and bust of commodity prices in 2006-2008, policymakers in many countries are debating on whether to impose tighter limits on positions taken by financial investors in commodities markets. It is important to interpret the two sides of our findings.
- On one hand, given the previous segmentation of commodities markets, the increasing presence of commodity index investors is likely to improve sharing of commodity price risk.
This means lower risk premia and thus higher prices on average for farmers and producers to sell their commodities.
- On the other hand, their presence also introduces a channel to spill over volatility from outside to commodities markets and across different commodities.
This trade-off requires a thorough examination. Before researchers can develop a reliable measure of the net effect, policymakers need to be cautious about imposing any stringent position limits on financial investors, as such limits also constrain the potential risk-sharing benefit.
Our findings also provide the basis for another more modest policy proposal. From our communications with investment advisers to various endowments and asset management funds, many of them do not fully realise the large increases in return correlations of commodities with each other and with stocks. Consistent with the volatility spillover effect, Figure 7 shows that the return correlation between the GSCI and S&P 500 stock index has also increased from around zero to above 0.5 in 2009.
Figure 7. Return correlation between GSCI and S&P stock indices
To the extent that the large inflow of commodity index investment is motivated by the low correlations observed in the historical data, many index investors might have overestimated the diversification benefit of investing in commodities. Thus, simply improving the public awareness of the increased correlations of commodities with each other and with stocks is likely to tame the rapid growth of commodity index investment and reduce the adverse volatility spillover effect.
References
Barberis, Nicholas and Andrei Shleifer (2003), âStyle investingâ, Journal of Financial Economics, 68:161-199.
Barberis, Nicholas, Andrei Shleifer, and Jeffrey Wurgler (2005), âComovementâ, Journal of Financial Economics 75, 283-317.
Bessembinder, Hendrik (1992), âSystematic risk, hedging pressure, and risk premiums in futures marketsâ, Review of Financial Studies 4, 637-667.
CFTC (2008), âStaff Report on Commodity Swap Dealers & Index Traders with Commission Recommendations!â
de Roon, Frans, Theo Nijman, and Chris Veld (2000), âHedging pressure effects in futures marketsâ, Journal of Finance, 55:1437-1456.
Erb, Claude and Campbell Harvey (2006), âThe strategic and tactical value of commodity futuresâ, Financial Analysts Journal 62 (2), 69-97.
Gorton, Gary and Geert Rouwenhorst (2006), âFacts and fantasies about commodity futuresâ, Financial Analysts Journal 62(2):47-68.
Hamilton, James (2009), âCauses and consequences of the oil shock of 2007-2008â, Working paper, UC San Diego.
Kilian, Lutz (2009), âNot all oil price shocks are alike: Disentangling demand and supply shocks in the crude oil marketâ, American Economic Review, 99:1053-1069.
Krugman, Paul (2008), âFuels on the hillâ, The New York Times, 27 June.
Kyle, Albert and Wei Xiong (2001), âContagion as a wealth effectâ, Journal of Finance, 56:1401-1440.
Masters, Michael (2008), âTestimony before the Committee on Homeland Security and Governmental Affairsâ, US Senate, 20 May.
Tang, Ke and Wei Xiong (2010), âIndex investment and financialization of commoditiesâ, NBER Working Paper 16385.
US Senate Permanent Subcommittee on Investigations (2009), âExcessive speculation in the wheat marketâ, Committee on Homeland Security and Governmental Affairs, 24 June.
