Does It Matter Who Trades Commodity Futures? Michel A. Robe

advertisement
Does It Matter Who Trades
Commodity Futures?
Michel A. Robe
Kogod School of Business
American University
Washington, DC
RFF, May 2012 – Michel A. Robe
Does It Matter Who Trades
Commodity Futures?
Bahattin
Büyükşahin*
*This
Michel
Robe*
presentation reflects the opinions of its authors only, and not those of the U.S. Commodities Futures
Trading Commission (CFTC), the International Energy Agency (IEA-OECD), any of the Commissioners, any
government, or the authors’ colleagues upon the staff of either institution.
RFF, May 2012 – Michel A. Robe
Commodity Trader Positions
and
Intra/Cross-Market Linkages
Bahattin
Büyükşahin*
Michael S.
Haigh
RFF, May 2012 – Michel A. Robe
Jeffrey H.
Harris
James A.
Overdahl
Michel A.
Robe*
I. Background
RFF, May 2012 – Michel A. Robe
Observations
 More investment money in commodity futures markets
 Thousands of hedge funds, commodity index funds, etc.
 Commodity assets under management (AUM):
broke the $400bn mark in 2011; inflows > $350bn in past 10 years (Barclays)
 What could this development mean for…
 Energy/Commodity Price Levels?
 ITF Report; Hamilton (2009); Kilian & Murphy (2011), Kilian & Hicks (2012)
 Commodity Market Volatility?
 Büyükşahin & Harris (2011); Brunetti et al (2011)
 Inter- or Cross-Market Linkages?
  My focus today
RFF, May 2012 – Michel A. Robe
RFF, May 2012 – Michel A. Robe
Year
5/5/2004
3/5/2004
1/5/2004
11/5/2003
9/5/2003
45
7/5/2003
5/5/2003
3/5/2003
1/5/2003
11/5/2002
9/5/2002
7/5/2002
5/5/2002
3/5/2002
1/5/2002
11/5/2001
9/5/2001
7/5/2001
5/5/2001
3/5/2001
1/5/2001
11/5/2000
9/5/2000
7/5/2000
Price (USD)
A. Intra-Market Linkages
Nearby, 1-yr and 2-yr WTI Oil Prices: 2000 to mid-2004
WTI Futures Prices (July 00 to June 04)
Nearby
1-year
2-year
30
15
RFF, May 2012 – Michel A. Robe
Year
7/6/2008
5/6/2008
3/6/2008
1/6/2008
11/6/2007
9/6/2007
7/6/2007
5/6/2007
3/6/2007
1/6/2007
11/6/2006
9/6/2006
7/6/2006
5/6/2006
3/6/2006
1/6/2006
11/6/2005
9/6/2005
7/6/2005
5/6/2005
3/6/2005
1/6/2005
11/6/2004
9/6/2004
7/6/2004
Price (USD)
Nearby, 1-yr and 2-yr WTI Oil Prices: mid-2004 to 2008
WTI Futures Prices (July 04 to August 2008)
145
Nearby
1-year
2-year
90
35
Nearby, 1-yr and 2-yr WTI Oil Prices: mid-2004 to 2008
150
Nearby WTI Price = p_0
1-yr out (p_12)
125
2-yr out (p_24)
100
75
50
25
0
1/4/2000
1/4/2001
1/4/2002
1/4/2003
1/4/2004
RFF, May 2012 – Michel A. Robe
1/4/2005
1/4/2006
1/4/2007
1/4/2008
1/4/2009
1/4/2010
1/4/2011
B. Cross-Market Linkages
“As more money has chased (...) risky assets,
correlations have risen. By the same logic, at
moments when investors become risk-averse and want
to cut their positions, these asset classes tend to fall
together. The effect can be particularly dramatic if
the asset classes are small—as in commodities. (...)
This marching-in-step has been described (...) as a
‘market of one’.”
The Economist, March 8, 2007.
RFF, May 2012 – Michel A. Robe
Equity & Commodity Indices, 1991-2009
500
400
300
200
100
Base: Mean 1991 = 100
0
92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08
S&P 500 Index (1991=100)
Dow Jones Industrial Index (1991=100)
S&P GS Commodity Total Return Index (1991=100)
DJ_AIG Commodity Total Return Index(1991=100)
RFF, May 2012 – Michel A. Robe
Post-Lehman?
275
L
e
h
m
a
n
GSCI (Commodity)
250
DJUBS (Commodity)
225
DJIA (US equities)
200
S&P 500 (US equities)
175
MSCI World Equities
150
125
100
75
50
Base: Jan. 2, 2001 = 100
25
RFF, May 2012 – Michel A. Robe
Questions for Today
 Describe today’s Energy & Commodity futures markets
 Maturity structure, Participant structure, Financialization
 Pricing Analysis
 Q1: Are near & far-month futures prices co-integrated?
 Q2: Do energy markets move in sync with other markets?
 Q3: What drives intra and/or cross-market linkages?
 Macroeconomic or market fundamentals?
 trading activity?
 both?
RFF, May 2012 – Michel A. Robe
1
II. Price Facts
RFF, May 2012 – Michel A. Robe
A. Intra-Market Evidence
Cross-Maturity (i.e., Intra-Market) Linkages
 Recursive co-integration analysis
 Idea: Notwithstanding shot-term shocks, is there a long-term
relationship between near-dated & far-dated futures prices?
 Sample: March 1989 to May 2011
 Prior to 2003, evidence is mixed
 up to 9 months cointegrated with nearby since mid-90’s
 BUT 1-yr and 2-yr out prices not cointegrated with nearby!
 After 2004, a different world?
 Yes – remains true post-Lehman
RFF, May 2012 – Michel A. Robe
Front-Back Price Cointegration, 1989-2011
6
Trace_0_1_2
Trace_0_6_9
5
Trace_0_12
Trace_0_24
4
Trace_0_12_24
3
March 2003
2
1
July 2004
0
RFF, May 2012 – Michel A. Robe
B. Cross-Markets Evidence
Commodity-Equity (i.e., Cross-Markets) Linkages
 Dynamic conditional correlation (DCC) analysis
 Idea: Account for time-variations in relationship & volatilities
 Sample period: January 1991 to May 2011
 Pre-Lehman, correlations between returns on passive commodity and
equity investments averaged about zero BUT fluctuated a lot
 True at daily, weekly & monthly frequencies
 True regardless of GSCI or DJ-AIG(UBS); S&P or DJIA
 Post-Lehman, a different world?
RFF, May 2012 – Michel A. Robe
 DCC estimates average close to Ø, fluctuate substantially over time
DCC_MR_GSENTR_SP
0.6
DCC_MR_GSENTR_MXWO
Egypt protests
0.4
0.2
0
-0.2
Lehman collapse
-0.4
RFF, May 2012 – Michel A. Robe
 Importance of accounting for volatility changes
 Rolling
vs.
 DCC
RFF, May 2012 – Michel A. Robe
III. Trading Facts
RFF, May 2012 – Michel A. Robe
A. Position Data – Crude Oil
 Publicly available data (COT, 2000-2011)
 Weekly CFTC Commitments of Traders (COT) Reports
 Highly aggregated
 All maturities are lumped together
 Traders grouped in 2 bins: “Commercial” vs. “Non-Commercial”
 vs. Our data: Large Trader Reporting System (LTRS)
 End-of-day positions of every individual large trader
 Non-public, CFTC only
 For every contract maturity
 Every day from July 2000 to March 2010
 Information on each trader’s business
RFF, May 2012 – Michel A. Robe
Our Detailed Data: WTI Sub-Categories
 Non-commercials

Hedge Funds (includes Commodity Pool Operators (CPOs), Commodity
Trading Advisors (CTAs), Associated Persons who control customer accounts,
and other Managed Money traders)
 Floor Brokers & Traders
 Non-Registered Participants (Traders not registered under the Commodity
Exchange Act (CEA); category includes non-MMT financial traders)
 Commercials
 “Traditional Commercials”
 Producers
 Manufacturers (refiners, etc.)
 Dealers (energy wholesalers, exporter/importers, marketers, etc.)
 Commodity Swap Dealers (includes arbitrageurs and CITs)
RFF, May 2012 – Michel A. Robe
1. Differential Growth at Near/Far Ends
o E.g., 1-3 years OI now > 1-3 months OI back in 2000
Open Interest In WTI Crude Oil Futures and
Adjusted Option Positions
1,200,000
> 3 yrs
1,000,000
12-36 months
3-12 months
< 3 months
800,000
600,000
400,000
200,000
< 3 months
3-12 months
12-36 months
> 3 yrs
0
2000 2001
2002 2003
2004 2005
2006 2007
RFF, May 2012 – Michel A. Robe
2008
2. “Financialization”
o Hedge Funds & Swap Dealers (incl. CITs) are up
RFF, May 2012 – Michel A. Robe
3. Heterogeneity within Broad Categories
o Good idea to break out Swap Dealers & Hedge Funds (CFTC 2009)
RFF, May 2012 – Michel A. Robe
4. Heterogeneity Extends to Options
o E.g., much lower net exposure of Swap Dealers & FBTs
RFF, May 2012 – Michel A. Robe
5. Differential Behaviors at Near/Far End
o E.g., Swap Dealers: net long “nearby” / net short “backdated”
RFF, May 2012 – Michel A. Robe
Nice Data – Show us it matters!
2010 OPEC
“observation
”:
Strong
positive
correlation
between net
hedge fund
positions and
crude oil
prices
RFF, May 2012 – Michel A. Robe
RFF, May 2012 – Michel A. Robe
RFF, May 2012 – Michel A. Robe
B. Generalizing to all GSCI Commodities
 We would like
 Detailed position data for all indexed & non-indexed commodities
 Bad news
 Some of the contracts are non-US  no data (e.g., Gas oil; Brent)
 Position data for RBOB gasoline are available only after 2006
 Good news
 We have data for 17 US commodity futures markets
RFF, May 2012 – Michel A. Robe
Main findings
 Financialization is a broad phenomenon
 Overall speculation is up
 From 10% excess spec till 2002 to 35-50% after 2005
 Commodity Index Trading is Up
 Swap Dealer positions account for about 35% of futures OI
 Hedge Funds are Up
 From 5-10% of the futures OI till 2002 to 25-30% after 2005
 Cross-Market Trading is Up
 Tripled since 2002
 Lengthening of the maturity curve is commodity-specific
 Energy 10 years ago  Metals 5 years ago  Ags today?
RFF, May 2012 – Michel A. Robe
IV. Does Trading Strengthen Linkages?
RFF, May 2012 – Michel A. Robe
A. Should It Matter Who Trades?
 Should trader identity matter for asset pricing?
 Theoretical reasons to believe trader identity matters
 Models show that less-constrained traders link asset markets
 During financial stress periods, contagion or retrenchment?
 Who is a “candidate” for enhancing linkages?
 Traditional commodity users, etc.?  Unlikely
 Index traders? Only insofar as they provide liquidity
 Hedge funds?  More likely
 Seek to exploit perceived mis-pricing
 Levered/subject to borrowing limits/wealth effects under stress
RFF, May 2012 – Michel A. Robe
B. Controls?
 1. Macroeconomy
 World vs. US cycles
 2. Commodity-specific developments
 E.g., production constraints
 3. Financial stress
 For many asset classes, co-movements increase under stress
 Do stress and some trading activity interact?
RFF, May 2012 – Michel A. Robe
C. Methodology
 ARDL regressions
 Stationarity
 Endogeneity
RFF, May 2012 – Michel A. Robe
V. Conclusions
RFF, May 2012 – Michel A. Robe
Bottom Line
 Impact of financial activity in commodity markets
 “Value arbitrageurs” help link markets
 Liquidity trading (or, conversely, market stress) matters, too
 Interaction – contagion through wealth effects?
 Information on OI composition is payoff-relevant
 2009 CFTC decision to disaggregate more
 Going forward
 Does speculation or hedging pressure drive commodity returns?
 Revisit commodity risk premia using disaggregated data
RFF, May 2012 – Michel A. Robe
Download