Crude oil futures markets: Does NYMEX trade and the de-linking
28
Professor Ronald D. Ripple, PhD Director, CREME Curtin University 19 September 2012 Crude oil futures markets: Does NYMEX trade and the de-linking of WTI and Brent affect ICE trade?
Transcript of Crude oil futures markets: Does NYMEX trade and the de-linking
Professor Ronald D. Ripple, PhD Director, CREME Curtin University
19 September 2012
Crude oil futures markets: Does NYMEX trade and the de-linking of
WTI and Brent affect ICE trade?
Crude oil futures markets
• Introduction
• Motivation
• Literature
• Methodology
• Data
• Results
• Conclusion and Discussion
Introduction
• This paper examines the relations between the crude-oil futures trading activities on the New York Mercantile Exchange (NYMEX) (now part of the Chicago Mercantile Exchange (CME) Group) and the Intercontinental Exchange (ICE; of London) exchanges, placed in the context of the apparent de-linking of WTI and Brent.
• The paper analyses the activities of NYMEX crude oil traders according to their open interest positions as reported in both the Legacy Commitment of Trader (LCOT) reports and the new Disaggregated COT (DCOT).
• The paper also analyzes the trading patterns of ICE-Brent traders, albeit for a much shorter time period.
• In addition, it addresses some questions regarding suggestions of “excessive” trading in crude oil futures contracts.
Motivation
• The motivation for this paper is to extend the literature analysing the relations between derivatives market traders and prices and to extend the understanding of the relations among the traders.
• A secondary motivation is to enhance the understanding of what the publically available data on trader activity may be able to reveal about these activities, and to some extent to identify the limitations and what they cannot say.
• The desire is that this will improve the understanding of the information upon which regulatory proposals are based.
Literature
• Haigh, M. S., Hranaiova, J., and Overdahl, J. (2005)
• Cooper, M. (2006)
• Ripple, R.D. (2008a; June)
• CFTC (2008; September)
• Ripple, R.D. (2008b)
Methodology
• For the most part, these preliminary analyses are conducted through graphical and tabular analyses.
• Some correlation analyses are undertaken, and since these are done on changes in variables there should be little concern with unit roots and spurious correlation; however, this aspect of the analyses nonetheless will be formally evaluated in future research.
Data • The paper examines crude oil prices, futures contract open interest,
and futures contract trading volume. • The primary price data employed are spot prices for West Texas
Intermediate (WTI) and Brent crude oil. These are drawn from the EIA webpage. Spot prices are employed in this preliminary analysis because the differences in trading days between the two exchanges makes matching futures settlement prices difficult; this will be worked on in the future.
• Futures contract open interest and trading volume data are drawn from the CRB Database for the NYMEX contract and from the ICE online database for the ICE-Brent contract.
• Open interest by trader category is drawn from the CFTC for the NYMEX contract (both LCOT and DCOT data) and from the ICE online database for the ICE-Brent contract (equivalent to DCOT).
• Some series date back to January 2000, while others date to January 2002. The ICE-Brent trader category open interest dates only to January 2011. All series come forward to June 2012.
The de-linking of WTI and Brent
• The de-linking appears to have taken force on or after 28 October 2010.
• Prior to this, while WTI typically traded at a premium to Brent, the price series can be seen to have crossed numerous times over the ten-year period.
• Since 28 October 2010, Brent has continually held a premium position with respect to WTI, but that premium has not been stable.
0
20
40
60
80
100
120
140
160
Jan
02
, 20
02
Ap
r 0
2, 2
00
2
Jul 0
2, 2
00
2
Oct
02
, 20
02
Jan
02
, 20
03
Ap
r 0
2, 2
00
3
Jul 0
2, 2
00
3
Oct
02
, 20
03
Jan
02
, 20
04
Ap
r 0
2, 2
00
4
Jul 0
2, 2
00
4
Oct
02
, 20
04
Jan
02
, 20
05
Ap
r 0
2, 2
00
5
Jul 0
2, 2
00
5
Oct
02
, 20
05
Jan
02
, 20
06
Ap
r 0
2, 2
00
6
Jul 0
2, 2
00
6
Oct
02
, 20
06
Jan
02
, 20
07
Ap
r 0
2, 2
00
7
Jul 0
2, 2
00
7
Oct
02
, 20
07
Jan
02
, 20
08
Ap
r 0
2, 2
00
8
Jul 0
2, 2
00
8
Oct
02
, 20
08
Jan
02
, 20
09
Ap
r 0
2, 2
00
9
Jul 0
2, 2
00
9
Oct
02
, 20
09
Jan
02
, 20
10
Ap
r 0
2, 2
01
0
Jul 0
2, 2
01
0
Oct
02
, 20
10
Jan
02
, 20
11
Ap
r 0
2, 2
01
1
Jul 0
2, 2
01
1
Oct
02
, 20
11
Jan
02
, 20
12
Ap
r 0
2, 2
01
2
WTI and Brent spot pricesDaily: 2 Jan 2002 - 30 Jun 2012
(source: EIA; US$ per barrel)
WTI
Brent
Spot price characteristics
• The prices for the two crude oils have actually reversed premium (discount) status reasonably frequently until 28 October 2010.
• With the exception of the mid-2008 period, the volatility for each price series appears relatively consistent.
-40
-30
-20
-10
0
10
20
30
Jan
02
, 20
02
Ap
r 0
2, 2
00
2
Jul 0
2, 2
00
2
Oct
02
, 20
02
Jan
02
, 20
03
Ap
r 0
2, 2
00
3
Jul 0
2, 2
00
3
Oct
02
, 20
03
Jan
02
, 20
04
Ap
r 0
2, 2
00
4
Jul 0
2, 2
00
4
Oct
02
, 20
04
Jan
02
, 20
05
Ap
r 0
2, 2
00
5
Jul 0
2, 2
00
5
Oct
02
, 20
05
Jan
02
, 20
06
Ap
r 0
2, 2
00
6
Jul 0
2, 2
00
6
Oct
02
, 20
06
Jan
02
, 20
07
Ap
r 0
2, 2
00
7
Jul 0
2, 2
00
7
Oct
02
, 20
07
Jan
02
, 20
08
Ap
r 0
2, 2
00
8
Jul 0
2, 2
00
8
Oct
02
, 20
08
Jan
02
, 20
09
Ap
r 0
2, 2
00
9
Jul 0
2, 2
00
9
Oct
02
, 20
09
Jan
02
, 20
10
Ap
r 0
2, 2
01
0
Jul 0
2, 2
01
0
Oct
02
, 20
10
Jan
02
, 20
11
Ap
r 0
2, 2
01
1
Jul 0
2, 2
01
1
Oct
02
, 20
11
Jan
02
, 20
12
Ap
r 0
2, 2
01
2
WTI minus BrentDaily: 2 Jan 2002 - 30 June 2012
(US$ per barrel: negatives imply WTI < Brent)
Period Average Difference
Full period -$1.123Through 27 Oct 2010 $1.41528 Oct 2010 onward -$14.567
Some apparent cyclicality in the premium/discount status beginning as early mid-2005.
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
Jan
30
, 20
02
May
30
, 20
02
Sep
30
, 20
02
Jan
31
, 20
03
May
31
, 20
03
Sep
30
, 20
03
Jan
31
, 20
04
May
31
, 20
04
Sep
30
, 20
04
Jan
31
, 20
05
May
31
, 20
05
Sep
30
, 20
05
Jan
31
, 20
06
May
31
, 20
06
Sep
30
, 20
06
Jan
31
, 20
07
May
31
, 20
07
Sep
30
, 20
07
Jan
31
, 20
08
May
31
, 20
08
Sep
30
, 20
08
Jan
31
, 20
09
May
31
, 20
09
Sep
30
, 20
09
Jan
31
, 20
10
May
31
, 20
10
Sep
30
, 20
10
Jan
31
, 20
11
May
31
, 20
11
Sep
30
, 20
11
Jan
31
, 20
12
May
31
, 20
12
WTI and Brent spot price - Coefficient of Variation20-day moving average
2 Jan 2002 - 30 June 2012
WTI
Brent
Linear (WTI)
Linear (Brent)
Price volatility has been relatively flat for each series with the exception of the 2008 period. And, Brent volatility appears to have “peaked” prior to that for WTI in the 2008 period.
Futures contract open interest
• What is open interest?
– It is the number of contracts at the end of the day that are still open for trading when the market reopens.
– There is no comparable measure in equities, since, unlike shares outstanding, open interest varies over the life of a contract.
– It is typically seen to provide a measure of market depth or liquidity.
– Importantly, there is no one-to-one relation between open interest and trading volume; they each provide different information regarding trading activity.
0
500,000
1,000,000
1,500,000
2,000,000
2,500,000
3,000,000
3,500,000
8/01/2002 8/01/2003 8/01/2004 8/01/2005 8/01/2006 8/01/2007 8/01/2008 8/01/2009 8/01/2010 8/01/2011 8/01/2012
Combined Open InterestNYMEX and Brent-ICE
0.00
200,000.00
400,000.00
600,000.00
800,000.00
1,000,000.00
1,200,000.00
1,400,000.00
1,600,000.00
1,800,000.00
8/01/2002 8/01/2003 8/01/2004 8/01/2005 8/01/2006 8/01/2007 8/01/2008 8/01/2009 8/01/2010 8/01/2011 8/01/2012
Total Open Interest positions: TuesdaysNYMEX-Brent crude oil contracts
Jan 2002 - June 2012
NYMEX
Brent
The generally increasing open interest in each contract seems to be influenced by both macroeconomic trends and a sense of the riskiness of the market. Even as Brent trading volume exceeds that for the NYMEX contract, Brent’s open interest remains smaller.
We will see later that the significant drop in open interest was due primarily to commercial/hedger traders and not attributable to so-called speculators, as is typically claimed.
CFTC Open Interest by Trader Category
• The CFTC reports open interest positions by trader category: – The Legacy Commitment of Trader (LCOT) data
• Commercial
• Non-commercial
• Non-reporting
– The Disaggregated Commitment of Trader (DCOT) data (available only from June 2006) • Producers/Merchants
• Swap Dealers
• Money Managers (this is where hedge funds are counted)
• Other Reporting
• Non-reporting
0
50,000
100,000
150,000
200,000
250,000
300,000
350,000
400,000
450,000
500,000
13/06/2006 13/06/2007 13/06/2008 13/06/2009 13/06/2010 13/06/2011 13/06/2012
NYMEX Trader Long PositionsJune 2006 - Aug 2012
PML
SWL
MML
OL
NRL
0.00
100,000.00
200,000.00
300,000.00
400,000.00
500,000.00
600,000.00
700,000.00
800,000.00
900,000.00
1,000,000.00
8/01/2002 8/01/2003 8/01/2004 8/01/2005 8/01/2006 8/01/2007 8/01/2008 8/01/2009 8/01/2010 8/01/2011 8/01/2012
NYMEX Traders Long PositionsJan 2002 - June 2012
NCL
CL
NRL
Commercials equal the sum of Producers/Merchants and Swap Dealers. Non-commercials equal the sum of Money Mangers and Other Reporters. Non-reporters have open positions less than 350 contracts. The plunge in 2008 is dominated by the Producers/Merchants.
I am showing long positions only. Each trader category is also well represented on the short side of trading activity, but it seems to be the longs that draw the most attention. It is clear that the Commercials dominate the market.
ICE and NYMEX
• The ICE only began to make available open interest positions by trader category from January 2011, which are comparable to those released by the CFTC.
• This does not allow for a comparison of trader category activity before and after the apparent de-linking, but we may compare the trading patterns on the two exchanges.
0
100,000
200,000
300,000
400,000
500,000
600,000
Brent-ICE Trader Long PositionsJan 2011 - Aug 2012
PML
SWL
MML
OL
NRL
0
50,000
100,000
150,000
200,000
250,000
300,000
350,000
NYMEX Trader Long PositionsJan 2011 - Aug 2012
PML
SWL
MML
OL
NRL
In the March through May period Money Managers enter the NYMEX market long at the same time Producers/Merchants are leaving. Short positions for Producers/Merchants did not change much, so MMs were providing hedging support.
There is a much clearer pattern of relative trader activity on the ICE with Producers/Merchants dominating. However, the combined Commercial traders clearly dominate both exchanges during this period following the price de-linking.
Open Interest and De-linked Price • One would expect that the apparent de-linking of the prices
would have a discernible effect on the trading activity of the two markets.
• One way to examine this is to look to the correlation between the changes in the price differential and changes in the open interest on each exchange. We may expect that an increasing premium for Brent over WTI would have a significant positive correlation with ICE-Brent open interest changes and a significant negative correlation with NYMEX open interest changes.
• Also, the correlation between the changes in open interest on each exchange may provide some insights; for example if the NYMEX contract is de-linked from the rest of the world, we may expect to see trading activity leaving the NYMEX and transferring to the ICE. Hence, we would expect a significant negative correlation.
• Using Tuesday observations on spot price changes and open interest position changes:
• I find that the correlation between price differential changes and NYMEX open interest changes to be -0.11, and for ICE-Brent open interest it is -0.08. Neither suggests a strong influence, and the sign for Brent is not as expected.
• I find the correlation between changes in NYMEX open interest and the ICE-Brent open interest to equal 0.32. Again, not a strong influence, and not the expected sign.
• It is also interesting to note that after the de-linking of prices, the correlation between the changes in the two prices actually increased. For the period before 28 Oct 2010, the correlation between price changes was 0.59, but after that date it was 0.66.
Open Interest and De-linked Price
Crude oil trading volume on the NYMEX and the ICE
• Recently, the financial media have made a big deal about ICE Brent gaining trading-volume superiority over the NYMEX-light sweet contract; some suggesting it is a first.
• As noted earlier, the recent (and even earlier) ICE-Brent contract trading volume has exceeded that of the NYMEX contract even while having lower open interest; what does this have to say about relative “excessive” trading between the two exchanges?
0
200,000
400,000
600,000
800,000
1,000,000
1,200,000
1,400,000
1,600,000
NYMEX Light Sweet and ICE BrentTuesday Total Trading Volume
Jan 2000 - June 2012
NYMEX Vol
ICE Vol
-600,000
-400,000
-200,000
0
200,000
400,000
600,000
800,000
1,000,000
Trading Volume Difference: NYMEX minus ICETuesday ObservationsJan 2000 - June 2012
This observation appears quite unusual. The volume for ICE on this day, 12 June 2012, was 1,139,190, however, the day before the volume was 724,743 and the day after it was 743,234.
Crude oil trading volume, at least as observed on Tuesdays, tends to be somewhat volatile on both exchanges. However, it does appear that the NYMEX volatility began to increase sooner at about the time its volume surged in early 2007. ICE appears to have taken nearly five years to close the gap.
The newness of the ICE-Brent trading volume exceeding that of the NYMEX contract would appear to be new only to those who did not observe the market prior to 2007.
What is excessive trading?
• I am not aware of any theoretically based definition for excessive trading.
• We typically think that more trading is a good thing, since this tends to assist with well-founded price determination. So what determines when we have too much of a good thing?
• How may one then at least try to determine whether or not crude oil futures trading activity has changed or fluctuated in market-disruptive ways?
Measures of trading activity (1) • It is frequently asserted that the notional volume of crude oil
traded via futures contract trade vastly exceeds the physical volume of crude oil consumed.
• One such example is Cooper (2006), where in the context of natural gas it is claimed that futures trade amounted to as much as 30 times the physical usage.
• For crude oil, the multiple is frequently stated to exceed 15 or 20 times, while even the NYMEX has claimed its trade to exceed global physical volumes by a factor of three.
• None of these is correct.
• In my June, 2008 Oil and Gas Journal article I demonstrate that futures trade represents a fraction of the physical when its volumes are placed on an equal footing (apples-to-apples) with the physical typically used for comparison, i.e., daily usage.
• The primary problem is that the bases for futures volumes and physical volumes have not been equalized. The physical is stated in terms of barrels per day, but the futures volumes used reflect traded barrels for all open contracts, which for the NYMEX contract is currently through December 2020. It makes no economic sense (or any other kind I can think of) to compare over seven years worth of traded volume to a single days consumption.
• When we make the appropriate adjustment, we find that on average over the period for the NYMEX contract trading volume amounts to 34% of physical, while the ICE-Brent contract trade amounted to 16% of physical trade, using 18 million barrels per day for the physical usage.
Measures of trading activity (1) – cont.
• Trading volume should not be examined in a vacuum; this may work for equities, but that is because shares outstanding is relatively static.
• Open interest evolves over the life of a contract, and it has evolved over time.
• Trading volume should be assessed relative to open interest. • If speculators in the futures market rush in on a daily basis
trading in ways that move the market, but leave by the end of the day (much like day traders in equities), we would observe spikes in trading volume relative to open interest.
• Open interest is only assessed at the end of the day, so a trader may enter and leave the market and have not influence on the open interest positions, while definitely having an observable impact on trading volume.
• Thus, a measure of daily open interest divided by daily trading volume should provide information regarding trends in trading and unusual shifts and changes in trading activity.
Measures of trading activity (2)
0
2
4
6
8
10
12
14
16
4/01/2000 4/01/2001 4/01/2002 4/01/2003 4/01/2004 4/01/2005 4/01/2006 4/01/2007 4/01/2008 4/01/2009 4/01/2010 4/01/2011 4/01/2012
NYMEX light sweet Open Interest divided by Trading VolumeTuesday observationsJan 2000 - June 2012
0
2
4
6
8
10
12
14
16
4/01/2000 4/01/2001 4/01/2002 4/01/2003 4/01/2004 4/01/2005 4/01/2006 4/01/2007 4/01/2008 4/01/2009 4/01/2010 4/01/2011 4/01/2012
Brent-ICE Open Interest divided by Trading VolumeTuesday observationsJan 2000 - June 2012The big upward spikes reflect
holidays. The downward trend for both exchanges implies that there is more trading per open contract, but recall that trading volume is less than physical usage.
We are not able to say anything about who is doing the trading. The Haigh, et al. (2008) paper showed that Non-commercials tended to hold their positions longer than did the Commercials, but the available data does not allow such a determination here.
Conclusions and Discussion • The apparent de-linking of WTI and Brent does not appear to
have had discernible impacts on the relative trading activity between the two contracts. It also does not appear to have played a significant role in how the different trader categories have traded through the period. What changes have occurred will have to be explained by other drivers, perhaps more related to general macroeconomic variables and the on-and-off economic recovery, and the disparity of the recovery in different parts of the world.
• The growth in open interest and trading volume may also reflect increased sophistication on the part of market participants and an increased comfort level with these risk mitigation instruments.
• The analyses have also shown that there is no evidence to support the frequent claims of excessive trading activity on the crude oil futures markets, whether it be within the US on the CME/NYMEX or on the ICE for the Brent contract.
Thank you!
Questions/Comments?
CREME website: business.curtin.edu.au/creme Email: [email protected] Telephone: (61 8) 9266 3935
Perth, Western Australia
Prof. Ronald D. Ripple Curtin University
