Commodity Price Co-movements: what do they tell us?

by

Valeria Rolli

Servizio Studi e Relazioni Internazionali Banca d’Italia

“Food Price Volatility and the Role of Speculation”

FAO Headquarters, 6 July 2012

Outline

•Commodity prices and speculation: a link hard to capture

•What has been achieved by the G20 process

•Commodity market financialization: no comprehensive statistics

•Co-movements between commodities: what do they tell us?

Commodity prices and speculation: a link hard to capture

In theory, financial investments could have both positive (adding depth and liquidity) and negative

(destabilizing) effects on commodity markets (due to market imperfections such as

asymmetric information, herding behaviour and multiple believes)

In practice, however, the influence of “speculation” is very hard to detect from the empirical point

of view (no wonder academic research is quite inconclusive)

Report of the G20 Study Group on Commodities (under the chairmanship of Mr. Nakaso)

[Executive Summary, p. 6]:

“Assessments of the impact of financial investors on commodity prices remain inconclusive.

Large changes in physical supply and demand provide plausible explanations for commodity

price trends over the past several years and existing literature finds limited signs of investors

causing sustained deviations from ―fundamentals. At the same time there are views that

greater investor participation has at times affected commodity price volatility and correlations

between commodity and stock markets.”

Commodity prices and speculation: a link hard to capture

Why the empirical link is so difficult to analyze?

• No complete coverage of speculative positions; when available, data are typically of too

low frequency or not detailed enough (do not allow to distinguish investors’ motives)

• Data on commodity market fundamentals also incomplete (the unexplained residual,

not due to fundamentals, results very high). Problems with poor data on current

fundamentals (supply and inventories) and also with data capturing expected future

market developments (“news”)

• Impossibility to detect precisely the (causal) direction of the relation between

speculative investments and commodity prices (do speculators move prices or do they

react to price movements?)

• Even when commodity price drivers are correctly identified, the channels of

transmission are hard to distinguish. Example: A change in financial conditions (due to

an exogenous monetary policy move) tends to affect commodity prices through multiple

channels (changes in expected aggregate demand and inflation, in addition to higher

speculative hoarding demand).

What has been achieved by the G20 co-operative process?

G20 Agricultural Ministers "Action Plan on Food Price Volatility and

Agriculture" (June 2011)

Launch of the Agricultural Market Information System (AMIS), aiming to provide global

monthly data on production and stocks for wheat, corn, rice and soybeans.

G20 Finance Ministers and Central Bank Governors Meeting (Washington,

April 2011)

Call for enhanced transparency in both cash and derivatives markets (including Over-The-

Counter derivatives), as recommended by IOSCO

Call for stronger regulation and supervision on derivatives markets to address market abuses

and manipulation (including through position management powers)

What has been achieved by the G20 co-operative process?

Open issues:

• Counterproductive national policy intervention on agricultural markets (import trade

barriers, temporary export restraints, bio-fuel mandatory requirements) recognized but

not yet adequately addressed

• Large consensus on improving market transparency (soft regulation) in derivatives

markets (especially over-the-counter) but direct intervention by regulators is more

controversial

• Regulation pertains to the micro domain, it may not prevent cyclical aggregate financial

inflows into commodity markets

• Changes in global liquidity and risk premia affect commodity markets via multiple

channels (expected aggregate inflation and demand, possibly financial inflows). A wider

approach is needed, encompassing better macroeconomic management (governance

of global liquidity, international coordination of monetary and exchange rate policies).

Financialization in commodity markets: no comprehensive

statistics

Measures based on weight of financial investments

BIS quarterly data on:

a) number of commodity derivatives contracts on organized exchanges: substantial

increase since early 2000

b) notional amounts outstanding on over-the-counter (OTC) markets (reflecting both price

and volume changes): collapse during 2008 crisis and quite subdued afterwards

US CFTC (Commodity Futures Trading Commission) weekly positions in regulated

derivatives markets by type of investors: peak of pure financial activity (relatively high

for oil) in the second half of 2008, collapse during the crisis (especially for oil),

subsequent gradual recovery (assessment by the “T index”)

Assets under management by financial institutions provided monthly by Barclays Capital:

rapid recovery since mid-2010 (figures are however inflated by ETP in precious metals)

reflecting strong development of new instruments (exchange-traded products and

medium term notes) while investments in commodity indices have slowed down

Taken together, these sources confirm the increasing relevance of financial

investments since early 2000, due to the combined effects of

deregulation and financial innovation; new operators and instruments

have emerged.

BIS quarterly data on global commodity derivatives markets

Futures and options contracts outstanding on commodity

exchanges

(number of contracts in millions)

0

10

20

30

40

50

60

70

80

Mar

.199

5

Mar

.199

6

Mar

.199

7

Mar

.199

8

Mar

.199

9

Mar

.200

0

Mar

.200

1

Mar

.200

2

Mar

.200

3

Mar

.200

4

Mar

.200

5

Mar

.200

6

Mar

.200

7

Mar

.200

8

Mar

.200

9

Mar

.201

0

Mar

.201

1

Mar

.201

2

US markets Other markets source BIS

OTC- notional amounts outstanding

(billions of US dollars)

-

2.000

4.000

6.000

8.000

10.000

12.000

14.000

Jun.

2003

Dec

.200

3

Jun.

2004

Dec

.200

4

Jun.

2005

Dec

.200

5

Jun.

2006

Dec

.200

6

Jun.

2007

Dec

.200

7

Jun.

2008

Dec

.200

8

Jun.

2009

Dec

.200

9

Jun.

2010

Dec

.201

0

Jun.

2011

Dec

.201

1

Gold Other precious metals Other commoditiessource: BIS

US CFTC weekly positions in regulated derivatives markets,

Total and by Investors

Open Interest Positions in Commodity Derivatives

(futures and options) (1)

0

500000

1000000

1500000

2000000

2500000

3000000

3500000

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012

CORN COPPER OIL WHEAT

Source: CFTC, Commitment of Traders.

(1) Total number of outstanding contracts.

" T- Index" of the degree of financialization (1)

(weakly data, 1 month moving average)

1

1,2

1,4

1,6

1,8

2

2,2

2,4

2,6

2,8

3

09-2006 03-2007 09-2007 03-2008 09-2008 03-2009 09-2009 03-2010 09-2010 03-2011 09-2011 03-2012

1

1,3

1,6

1,9

2,2

2,5

2,8

3,1

3,4

3,7

4

COPPER CORN WHEAT OIL (rhs)

Source :CFTC, Commitment of Traders.

(1)Number of contracts in which both counterparts are non commercials operators over the total

number of contracts held by commercials.

Barclays Capital estimates of assets under management by

financial institutions

Financialization in commodity markets: no comprehensive

statistics

Measures based on price correlations across different commodities, and

between commodities and other financial assets

• Correlations (co-movements) between different commodities and commodity indexes

have increased in importance, with a structural jump since March 2008

• Commodity and US equity indexes have also become positively correlated (2008 break

however not significant, due to a volatile relation across the whole period).

These trends have been taken as evidence of the influence of financial

investments (and the diffusion of commodity indices which bundle together

single products).

Instead, we look at a different explanation, based on the influence of an

underlying common factor (a common shock) affecting both equity

prices and commodity demand

Correlations between different commodity indexes and

between commodities and other financial assets

Pairwise correlations between main IMF commodity

indices

-0,4

-0,3

-0,2

-0,1

0,0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011

agriculture-metals agriculture-energy metals-energy

Source: Di Nino and Natoli, Bank of Italy.

Pairwise correlations between main commodity

indices and the SP500

-0,4

-0,3

-0,2

-0,1

0,0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011

agriculture energy metals

Co-movements between commodity prices:

what do they tell us?

• Colleagues at the Bank of Italy (joint work by Virginia Di Nino and Filippo

Natoli) have taken the monthly yields of the main IMF Commodity Price

Indices over Jan 00 – Feb 12 period and applied a principal component

analysis. They have found that the first component (so called “common

factor”) explains about 40% of the total variance in the Jan 00 – Mar 08

period, and up to 70% of the total variance in the Apr 08 – Feb 12 period.

This indicates that the common factor has become dominant in

explaining developments since the 2008 crisis.

• They have then investigated the reason behind the increased explanatory

power by the common factor. Possibility n.1: the elasticity between the

common factor and the original commodity indexes (technically speaking, the

factor loading coefficients) has changed. Possibility n.2: the variability of the

common factor has increased. The latter possibility has been confirmed

by the data. This indicates that a bigger common shock has occurred.

Principal component analysis applied to the monthly yields

of the main IMF commodity price indexes

Source: Di Nino and Natoli, Bank of Italy.

Jan 2000 - Mar 2008 Apr 2008 - Feb 2012

Food and beverages 0.43 0.49

Industrial agricultural 0.50 0.47

Metals 0.51 0.52

Oil 0.55 0.52

Share of Total Variance explained

by Principal Components First Principal Component Loading Factor

Co-movements between commodity prices:

what do they tell us?

• Finally, they have tried to give some economic interpretation to the Common

Factor, examining linkages with real and financial variables: world industrial

production (proxy for global demand); nominal effective exchange rate of the

US dollar; developments in the US equity index (S&P 500); 10 year─3 month

spreads in US interest rates (proxy for monetary policy).

• They have found that developments in EMEs’ industrial production and

the $ nominal effective exchange rate are significant in explaining the

movements in the Common Factor (while the other determinants are not

significant)

This indicates that:

• Global demand dynamics is significantly linked to commodity yields, with

developments in EMEs’ economic activity increasingly important.

• The (negative) influence of the dollar captures multiple (monetary, financial

and real) channels: a weakening $ may depend on looser US monetary

policy; may increase commodity demand in countries with appreciating

currencies (due to lower costs); may increase speculative demand for

commodity assets by financial investors (hedging against financial and

inflation risks associated with dollar depreciation).

Possible economic interpretation of the “Common Factor”

Source: Di Nino and Natoli, Bank of Italy.

50

65

80

95

110

125

140

155

170

185

200

gen-

00

lug-

00

gen-

01

lug-

01

gen-

02

lug-

02

gen-

03

lug-

03

gen-

04

lug-

04

gen-

05

lug-

05

gen-

06

lug-

06

gen-

07

lug-

07

gen-

08

lug-

08

gen-

09

lug-

09

gen-

10

lug-

10

gen-

11

lug-

11

gen-

12

-2

-1

0

1

2

3

4

5

6

7

8

IP adv. countries IP eme. countries

SP&500 Common Factor

NEER interest rate spread (rhs)

IMF Monthly Commodity Indexes Common Factor

and selected Real and Financial Variables

(2000 = 100)

Thank you for your

attention