Commodity Markets and Commodity Mutual Funds

Chris Plantier Investment Company Institute

October 15, 2012

Why Commodity Investment Doesnt Explain Commodity Price Increases

I. Commodity Investment Globally and the Financialization of Commodities

II. Spurious Relationship btw Assets and Prices

III. Economic Fundamentals Strong Global Growth and Weak U.S. Dollar
IV. An Empirical Horse Race

V. Why Commodity Mutual Funds Cant Explain Higher Commodity Prices

2

Commodity Investment Globally

According to BarCap, worldwide assets under

management in commodity investment products (ETPs, commodity swaps, and medium-term notes) stood at $426 billion in November 2011, compared to $156 billion in November 2008

Most of the increase ($170 billion) represents net

inflows from investors spread across a number of commodity markets; the remainder$100 billionreflects the recovery in commodity prices since late 2008
3

Global Commodity Markets are Massive With Over $10 Trillion Monthly Turnover

Barcap estimates global commodity futures and

options markets at $11.8 trillion in October 2011

However, size varies by commodity with futures

and options markets of gold, zinc, crude oil, and copper all exceeding $1 trillion

Global Commodity Market Size

Billions of dollars
Commodity West Texas Intermediate and Brent crude oil Live cattle CME Heating oil and gasoil Unleaded gasoline Gold Silver Zinc Copper Aluminium Corn Wheat CBOT Soybean Total sales in spot market $ Billions, annual 2,500 1,500 800 705 182 21 27 144 89 104 165 120 Trading volume in futures and options markets $ Billions, monthly 2,800 76 982 318 2,067 369 1,910 1,228 484 Futures and options market open interest $ Billions 279 21 74 33 145 30 13 58 47 64 17 55

291
61 375

Totals

6,357

10,964

835.0

Note: Spot (physical) market value is calculated using supply and average price for 2010 for each individual commodity. Source: Barclays Capital 5

Impact of Financialization of Commodities on Commodity Prices

Any price impact from index funds likely stems

from two sourcesnew flows and the rebalancing of positions over time

In theory, fund flows could impact prices as

some critics argue, empirical question though

Rebalancing generally stabilizes prices

P fund (or swap dealer) sells position and P fund (or swap dealer) buys position However, monthly roll may impact prices at certain times of the month
6

Financialization of Commodities, a.k.a. the Massive Passive Argument

Sometimes referred to as massive passives, commodity index funds have created a massive, ongoing demand for futures contracts unconnected to normal supply and demand for the underlying commodities... the more index funds and their swap dealers push to buy long future contracts and outnumber the speculators seeking to buy shorts, the more their buying pressure, by the very nature of supply and demand, will drive up the price of the long contracts. The resulting higher futures prices then translate all too often into higher prices for the underlying commodities, in part because so many of the contracts for the underlying commodities use futures prices as the commodity selling prices. Senator Levins Opening Statement Nov 2011

Commodity Prices Rose Over the Last Fifteen Years

Monthly, 1997-2011*
Price per unit (dollars) 2,000 1,800 1,600 Price per unit (dollars) 350
Gold (left scale), price per ounce

300 250
Corn (right scale), price per metric ton

1,400
1,200 1,000 800 600 400 200 0 1997
*Data to December 2011 Source: World Bank

200 150 100 50

WTI (right scale), crude oil price per barrel

0 1999 2001 2003 2005 2007 2009 2011

Assets and Number of Commodity Mutual Funds (CMFs)

Monthly, 2004-2011*
Assets (billions) $60 $50 Number of funds 30 25

$40
Number of funds (right scale) $30 $20 $10 $0 2004 Assets (left scale)

20
15 10 5 0 2005 2006 2007 2008 2009 2010 2011

*Data through December 2011 Source: Investment Company Institute 9

Commodity Mutual Fund (CMF) Assets and Commodity Price Indexes

Monthly, 2006-2011*
Assets (billions) $60 S&P GSCI (right scale) $50
Dow Jones-UBS Commodity Index (right scale)

Index level 200 175

$40
$30 $20 $10 $0 2006 2007 2008 2009 2010 2011 Commodity mutual fund assets (left scale)

150
125 100 75 50

*Data through December 2011 Note: Prices indexed to 100 in January 2006 Sources: Investment Company Institute and Bloomberg 10

Emerging Market Industrial Production Growth and Commodity Price Growth

Monthly, 2005-2011*
Percentage year over year 20% Percentage year over year 50% Emerging market industrial production growth (left scale) 40% 30% 20% 10% 5% 0% -5% -10% 2005 2006 2007 2008 2009 2010 2011 10%

15%

0%
-10% -20% Dow Jones-UBS Commodity Index (right scale) -30% -40% -50% -60%

*Data to October 2011. Note: The correlation between the two growth rates is 0.82. Sources: Netherlands Bureau for Economic Policy Analysis and Bloomberg 11

Commodity Prices & U.S. Dollar Value

Monthly, 2004-2011*
Commodity price index 900 800 700 S&P GSCI (left scale) 600 80 500 400 300 200 2004 Broad trade-weighted exchange value of U.S. dollar (right scale) 2005 2006 2007 2008 2009 2010 2011 85 90 95 Trade-weighted Index (inverted scale) 65 70 75

*Data to October 2011. Note: The correlation between the two series is -0.87. Sources: Bloomberg, Federal Reserve 12

Estimated Monthly/Weekly Regressions

Prices and contemporaneous CMF flows

(1) ln(Pt) = + *Ct/At-1,

Commodity prices, CMF flows, & fundamentals

(2) ln(Pt) = f(ln(Pt-1), Ct/At-1, ln(USDt),

ln(EMt)) Lead/Lag relationship (VAR) on weekly data

(3) ln(Pt) = f(ln(Pt-k), Ct-k/At-1-k), (4) Ct/At-1 = f(ln(Pt-k), Ct-k/At-1-k),

13

Monthly Results
Intercept ln(Pt-1) Ct/At-1 ln(USDt) ln(EMt) 0.2876** (0.1385) -2.5333*** (0.5625) 2.4242*** (0.8163) 0.0557 0.0456 1.738 2004M02 to 2011M12 0.2985 0.2832 1.946 2004M02 to 2011M12 0.1115 0.0920 2.106 2004M02 to 2011M11 Equation 1 Fund Flows, Ct -0.0062 (0.0093) Equation 2 USD only -0.0464 (0.5370) -0.1819 (0.1198) Equation 2 EM only -1.3131 (0.8889) -0.0296 (0.1044) Equation 2 All -1.8066** (0.7105) -0.3140*** (0.0836) 0.2609*** (0.0717) -2.4282*** (0.4500) 1.6917** (0.6472) 0.3955 0.3683 2.012 2004M02 to 2011M11 Equation 2 w/o Ct -1.1895* (0.6807) -0.2944*** (0.0969)

-2.3703*** (0.5285) 1.8948*** (0.6327) 0.3512 0.3296 1.995 2004M02 to 2011M11

R-squared Adj. R-squared D-W Statistic Sample

*denotes statistical significance at the 10 percent level ** denotes statistical significance at the 5 percent level *** denotes statistical significance at the 1 percent level
14

ForecastsEconomic Fundamentals versus Commodity Mutual Fund Flows

Commodity price index level, monthly, 2004-2011*
500 450 400 350 300 250 200 Forecast based only on commodity fund flows Dow Jones-UBS Commodity Index

150
100

Forecast based only on economic fundamentals

2004

2005

2006

2007

2008

2009

2010

2011

*Data and dynamic forecasts are from February 2004 to November 2011. Note: The correlation between the Dow Jones-UBS Commodity Index and the forecast based on economic fundamentals is 0.80. It is -0.05 for the forecast based on flows. Source: Bloomberg 15

Net New Cash Flow to CMFs and Commodity Price Changes

Monthly, 2004-2011*
Billions
$2.0

Monthly percent change

80%

$1.5

Monthly net new cash flow (left scale)

60% Dow Jones-UBS Commodity Total Return Index (right scale) 40%

$1.0

$0.5

20%

$0.0

0%

-$0.5 2004

-20% 2005 2006 2007 2008 2009 2010 2011

*Data as of December 2011 Sources: Investment Company Institute and Bloomberg (Dow Jones UBS Commodity Total Return Index) 16

Conclusion

Fundamentals, not funds, drive commodity

prices, e.g., EM growth and USD matter

Flows to commodity mutual funds have little or

no influence on commodity prices

This finding is consistent with academic

research showing that financialization has not driven commodity prices important benefits for investors diversification and natural hedge against inflation
17

Investing in commodity mutual funds provides

Why Commodity Mutual Funds Cant Explain Higher Commodity Prices

First, commodity mutual funds experienced net outflows
on average from January 2006 to June 2008 while commodity prices rose

Second, flows into commodity mutual funds are spread

across a wide range of markets and thus do not concentrate investment in a particular commodity (see Figure 15 in paper)

Finally, the $47.7 billion in commodity mutual funds as

of December 2011 is miniscule relative to the size of global commodity markets

18

Commodity Mutual Funds' Implied Position in Commodity Markets

Market
WTICrude Natural gas Gold Soybean Copper Corn Aluminium Wheat CBOT Heating oil Unleaded gasoline Live cattle CME Sugar Silver Soybean oil Zinc Coffee Nickel Lean hogs CME Cotton Brent crude Gasoil Wheat (KBOT) Lead Implied weight in individual commodities1 Percent 16.2% 10.4% 9.7% 7.3% 7.1% 6.8% 4.9% 4.5% 3.7% 3.6% 3.3% 3.2% 3.0% 2.6% 2.6% 2.2% 2.1% 1.9% 1.9% 1.7% 0.7% 0.1% 0.0% Implied dollar position in commodity, assets2 Billions $7.7 4.9 4.6 3.5 3.4 3.2 2.3 2.1 1.8 1.7 1.6 1.5 1.4 1.3 1.3 1.1 1.0 0.9 0.9 0.8 0.3 0.0 0.0 Share of index in market volume Percent 0.5% 1.5% 0.2% 0.9% 0.3% 1.1% 0.5% 3.5% 0.5% 0.5% 2.1% 1.8% 0.4% N/A 0.1% 1.3% 1.1% 2.3% 3.6% 0.1% 0.1% 0.1% 0.0%

Feeder cattle CME Cocoa Tin Palladium

0.0% 0.0% 0.0% 0.0%

0.0 0.0 0.0 0.0

0.2% 0.1% 0.0% 0.0%

19

Footnotes to Commodity Mutual Funds' Implied Position

Note: Based on December 2011 Assets of $47.7 billion
1 Implied

weight is calculated from the weights in the Dow-Jones UBS and S&P GS commodity indexes. Each is weighted according to the assets of commodity mutual funds tied to the underlying index. For example, over 90 percent of commodity mutual funds assets are linked to the Dow-Jones UBSCI and less than 10 percent to the S&P GSCI; each of which has a weight, respectively, of 29.9% and 14.7% on WTI oil. That implies that an average weight for commodity mutual funds of 16.2% as of December 2011. dollar position is the corresponding weight multiplied by total assets in commodity mutual funds as of December 2011 ($47.7 billion). For example, commodity mutual funds have an implied weight of 16.2% of their $47.7 billion in assets invested in WTI crude oil, for an estimated dollar position of $7.7 billion.

2 Implied

Sources: Dow Jones-UBS, Barclays Capital

20

Annex
Some literature on impact of commodity investment Figures 8 through 12, and Figure 14 from paper 8-global size of flows, 9-rebalancing and oil, 10examples and #s of commodity investment products, 11-index weights in two broad commodity indexes, 12CPI ex FE versus S&P500 and commodity index, and 14-weekly CMF flows since early 2009

Weekly regression results

21

Commodity index-linked new flows do not affect returns

Stoll and Whaley 2010 find that neither commodity
indexlinked flows nor monthly rolls cause futures price levels to change across a wide variety of commodity markets

Irwin and Sanders 2011b find little evidence that

index-linked investment affects commodity market returns or volatility

Using internal CFTC data, Aulerich, Irwin, and

Garcia 2010 find negligible evidence that daily index-linked investment affects commodity returns in 12 agricultural markets
22

Rebalancing impact on commodity price volatility

Irwin and Sanders 2010 suggests that commodity
index investors may reduce commodity price volatility because the indexes fixed weights force them to sell into markets with the greatest price increases and buy into markets with falling prices

Aulerich, Irwin, and Garcia 2010 shows that index

investors can dampen volatility in some markets

Brunetti, Byksahin, and Harris 2011 examine

daily swap dealer positions (a proxy for index investment) and find no evidence that these positions contribute systematically to price changes or volatility in the crude oil, natural gas, and corn
23

Monthly rolls of positions do not cause futures price levels to change

Mou 2010 and Frenk and Turbeville 2011 examine the
period when index investors typically exit futures positions and roll into new positions. They find that the spread between prices for nearby and next-nearby widens during the roll, but no persistent levels effect

Stoll and Whaley 2010 find that monthly rolls do not

cause futures price levels to change across a wide variety of commodity markets

Kastner 2010 shows that United States Natural Gas (a

commodity ETF) has no significant effect during the time the monthly roll occurs
24

8: Investment Flow into Global Commodity Markets by Sector, 12MSum

Billions of dollars, monthly, 20092011*
$35

Energy
$30 $25 $20 $15 $10 Agriculture Precious metals

$5 Base metals
$0 2009
*Data to November 2011. Source: Barclays Capital
25

2010

2011

9: Oil Price Versus Futures Equivalent Position of Commodity Index Traders

Daily, December 2007 to December 2008
Price per barrel 160 140 120 100 80 60 40 20 0 12/3/2007 Futures equivalent (right scale) 320 Oil Price (left scale)

Number of contracts (thousands)

420

400

380

360

340

300 3/3/2008 6/3/2008 9/3/2008 12/3/2008

Sources: Federal Reserve, U.S. Commodity Futures Trading Commission 26

10: Number of Commodity ExchangeTraded Products and Mutual Funds

Commodity ETFs Commodity ETNs Commodity mutual funds1 Managed futures strategy mutual funds2

December 1, 2004 December 1, 2006

1 (0) 6 (2)

0 (0) 3 (2)

2 7

0 0 2 13 20

December 1, 2008
December 1, 2010

18 (3)
28 (4) 34 (4)

42 (6)
43 (7) 61 (8)

12
23 30

September 30, 2011

1 Commodity

mutual funds are mutual funds whose primary investment objective is to give investors broad exposure to commodities by benchmarking to commodity indexes that are diversified across a wide array of commodities. 2 Managed futures strategy mutual funds are those that seek to give investors exposure to commodities, interest rates, and exchange rates through derivatives such as futures and swaps. To date, these funds have not been predominately invested in commodities, so are included in this table purely for completeness. Note: Number in parentheses denotes number of broad-based commodity ETFs or ETNs. Source: Morningstar 27

11: Commodity Index Weights, 2011

Commodity WTIcrude Brent crude Gasoil Heating oil Corn Unleaded gasoline Copper Wheat CBOT Gold Natural gas Soybean Live cattle CME Aluminium Sugar Lean hogs CME Cotton Coffee Wheat (KBOT) Nickel Silver Zinc Lead Feeder cattle CME Cocoa Soybean oil Tin Palladium S&P GSCI 29.9% 16.8% 7.2% 5.3% 5.1% 4.9% 3.6% 3.4% 3.4% 2.8% 2.7% 2.5% 2.4% 2.4% 1.4% 1.3% 1.1% 0.9% 0.7% 0.6% 0.6% 0.5% 0.4% 0.3% 0.0% 0.0% 0.0% Dow Jones-UBS Commodity Index 14.7% 0.0% 0.0% 3.6% 7.0% 3.5% 7.5% 4.6% 10.5% 11.2% 7.9% 3.4% 5.2% 3.3% 2.0% 2.0% 2.4% 0.0% 2.3% 3.3% 2.9% 0.0% 0.0% 0.0% 2.9% 0.0% 0.0%
28

Note: Weights on tin and palladium are zero in both indexes. Sources: Dow JonesUBS, Barclays Capital

12: Commodity Index and S&P 500 index vs. CPI F&E Component
Dollars invested, monthly, 2002-2011*
$35,000 $30,000 Dow Jones-UBS Commodity Total Return Index

$25,000
$20,000 $15,000 $10,000 $5,000 $0 2002 S&P 500 index (total return )

Consumer Price Index, food and energy component

2003

2004

2005

2006

2007

2008

2009

2010

2011

*Data are as of December 2011 Sources: Bloomberg, U.S. Bureau of Labor Statistics 29

14: Net New Cash Flow to CMFs and Weekly Price Changes
Weekly, 2009-2011*
Billions
$2.0

Weekly percent change

40%

$1.5

Weekly net new cash flow (left scale)

30% Dow Jones-UBS Commodity Total Return Index (right scale) 20%

$1.0

$0.5

10%

$0.0

0%

-$0.5 2009

-10% 2010 2011

*Data as of December 2011 Sources: Investment Company Institute and Bloomberg (Dow Jones UBS Commodity Total Return Index ) 30

Weekly Contemporaneous Results

Equation 1 -0.00157 (0.00173) 0.2776** (0.1188) 0.0195 0.0171 2.002 1/12/2004 to 12/26/2011 Equation 1 with S&P GSCI -0.00217 (0.00243) 0.3209** (0.1611) 0.0138 0.0114 1.996 1/12/2004 to 12/26/2011

, Intercept , Slope R-squared Adjusted R-squared Durbin-Watson Statistic Sample

*denotes statistical significance at the 10 percent level ** denotes statistical significance at the 5 percent level *** denotes statistical significance at the 1 percent level

31

Weekly VAR Results

Intercept Coef on ln(Pt-1) ln(Pt-2) ln(Pt-3) Coef on Ct-1/At-2 Ct-2/At-3 Ct-3/At-4 R-squared Adjusted R-squared Sample ln(Pt) -0.00008 (0.0016) 0.0016 (0.0502) -0.0530 (0.0503) 0.0334 (0.0500) -0.0200 (0.1656) 0.3049* (0.1606) -0.1482 (0.1516) 0.0125 -0.0021 2/02/2004 12/26/2011 Ct/At-1 0.0012** (0.0005) 0.0189 (0.0149) -0.0121 (0.0149) -0.0018 (0.0148) 0.2474*** (0.0490) 0.3106*** (0.0475) 0.2172*** (0.0448) 0.5515 0.5448 2/02/2004 12/26/2011

*denotes statistical significance at the 10 percent level ** denotes statistical significance at the 5 percent level *** denotes statistical significance at the 1 percent level
32