The Great Debate

Don’t bank on return of backwardation

May 20, 2010

Many energy analysts are predicting the crude market will move into backwardation before the end of the year.

Increasing demand and rising refinery runs will, in their view, reverse the unusual build up of inventories around the NYMEX delivery point at Cushing, and the market should revert to a more normal term structure.

The extreme contango visible at the front end of the NYMEX futures curve in the last seven weeks is certainly evidence of a “dislocation” caused by congestion around the delivery point. Front-month NYMEX futures have been trading at abnormally large discounts not only to second- and third-month NYMEX futures but also to Brent and other spot crudes such as Tapis.

Recent upticks in refinery runs, coupled with the forthcoming driving season, and continued economic recovery, have the potential to tighten the markets for crude oil and refined products over the summer. The question is whether this will simply narrow the contango from its current extreme level or push the market level or even into a backwardation.
Recent experience suggests the contango is set to narrow, but will not disappear entirely. Contango, not backwardation, is the “new normal” in oil markets.

Before 2005, WTI crude prices were more often in backwardation than contango. Backwardations were both larger and more frequent than contango (Charts 1-4).
(1) http://graphics.thomsonreuters.com/ce/CL-STRUC-1.pdf

(2) http://graphics.thomsonreuters.com/ce/CL-STRUC-2.pdf

(3) http://graphics.thomsonreuters.com/ce/CL-STRUC-3.pdf

(4) http://graphics.thomsonreuters.com/ce/CL-STRUC-4.pdf

In the 15 years from 1990 to the end of 2004, the front six months of the crude oil curve traded in backwardation for 127 months out of a total of 180 (70 percent of the time). The half-year backwardation averaged 2.0 percent of the flat price in 1990-1994, rising to 3.6 percent in 1995-1999 and 8.6 percent in 2000-2004.

From 2005 onwards, however, the trading pattern has inverted completely. The front six months of the curve have been in backwardation just 13 months out of a total of 68 (19 percent of the time). The half-year timespread has traded in an average contango (not backwardation) of 3.9 percent.

Part of the shift stems from the global recession. But it is worth noting that the market has traded in contango during all phases of the cycle: strong growth and rising prices (2007-H1 2008), slump (H2 2008-H1 2009) and recovery (H2 2009-H1 2010). In fact the only period of sustained backwardation trading was the 11 months between August 2007 and May 2008.

The shift in term structure is not confined to the first six months of the forward price curve. It is also evident in the second six months and at points further in the future.


It is tempting to blame the shift from backwardation to contango on well-known delivery problems with the NYMEX crude contract. The landlocked delivery point at Cushing with its limited storage and congested pipeline infrastructure has increasingly been left behind as the physical market has moved to the U.S. Gulf Coast and re-oriented towards imports of seaborne global crude.

But a narrow focus on the idiosyncratic qualities of WTI misses the more profound shift that has been occurring. The same adjustment in term structure from backwardation to contango, or from a narrow contango to a wider one, is evident across a wide range of major commodity contracts (Charts 5-7).

(5) http://graphics.thomsonreuters.com/ce/GSCI-TERM.pdf

(6) http://graphics.thomsonreuters.com/ce/GSNE-TERM.pdf

(7) http://graphics.thomsonreuters.com/ce/GSEN-TERM.pdf

Until 1991, the whole basket of commodity contracts included in the Standard and Poor’s Goldman Sachs Commodity Index traded mostly in backwardation, producing a positive yield for investors as long positions in maturing futures contracts were rolled forward each month.

For the rest of the 1990s and the first years of the last decade, market structure across the commodities included in the basket was roughly balanced between backwardations and contangos. On average roll returns were flat or slightly negative, indicating a slight bias towards contango had emerged.

But since late 2004, the commodities in the index have traded in a very large contango on average, inflicting significant and sustained negative roll returns on long investors. In fact, the index has generated positive roll yield for investors in only 3 out of 70 months since August 2004 (January 2008, April 2008 and May 2008).

Again, it is tempting to attribute the shift from backwardations/positive roll yields to contangos/negative rolls as mostly about the changing behaviour of the oil market. But it would not be true.


The GSCI’s high-weighting towards crude oil (WTI and Brent) is well known. Index operators offer several variants which reduce that weighting in order to improve the index’s diversification properties.

But the shift in term structure is not wholly or mainly due to WTI. The same shift is evident in the non-energy components included in the GSCI non-energy sub-index as well as the energy ones.

Non-energy contracts shifted from backwardation to contango trading from the late 1990s onward, five years ahead of the shift for energy contracts, where the shift did not occur until late 2004. There is no meaningful difference in how often non-energy contracts trade in backwardation (10 months out of the last 70) compared with their energy counterparts (5 months out of 70).

In fact, it could be argued that positive yields from backwardations in oil and other energy futures helped disguise deteriorating roll yields from farm contracts, industrial metals and precious metals in the late 1990s and early 2000s, until the oil market itself flipped into contango after 2004.


Even a casual analysis of the returns on the GSCI and its components, as well as forward curves for major futures contracts, shows a profound, seemingly enduring shift. There is less agreement on what has caused it, and whether it will persist in future.

Altered curve structures must be the result of some combination of internal changes intrinsic to each market, external influences affecting all markets, or both. It is hard to see what internal developments could have caused such a substantial shift across so many very different markets at roughly the same time. So it is more likely that an external factor is at work.

The most plausible candidate is the increase in trading and liquidity (as well as positive hedge pressure) associated with the large-scale entry of index funds and other new participants into commodity markets, especially from 2004 onwards, as investors showed increased interest in commodity futures and options as an “asset class”.

Investor longs and the structural contango are also consistent with the commodity markets’ apparent capacity to carry much higher level of inventories than before without triggering a collapse in spot prices, for example during 2009 and early 2010.

If this explanation is true, and I have not seen a better alternative yet, it suggests the entry of substantial numbers of investors has forced a structural shift in pricing. Other things equal, it should ensure the oil market, and other markets, continue to trade much more in contango than in the past.

It does not rule out a return of backwardation in the oil market, especially if the physical market eventually tightens. But a backwardation later this year seems unlikely, based on recent trends, and if it occurs is likely to be fleeting.

Post Your Comment

We welcome comments that advance the story through relevant opinion, anecdotes, links and data. If you see a comment that you believe is irrelevant or inappropriate, you can flag it to our editors by using the report abuse links. Views expressed in the comments do not represent those of Reuters. For more information on our comment policy, see http://blogs.reuters.com/fulldisclosure/2010/09/27/toward-a-more-thoughtful-conversation-on-stories/