The Commodity Contango Conundrum & Exchange Traded Funds

ABSR Research LogoIf your ETF or ETN holds a single commodity or basket of commodities via exchange traded futures contracts, contango, is not your friend. Historically commodity contract curves have moved back and forth from periods of backwardation and contango with regular frequency. Our research suggests however, that markets are trending toward a more persistent contango term structure. This change while broad based is particularly evident in energy contacts. Commodity ETFs that are single commodity or broad based may be at particular risk depending on the contracts they hold. Despite this contango conundrum, there are methods to potentially mitigate the negative impact of contango on long only returns.

The Forward Curve

To fully understand how the structure of the forward curve effects long only performance and clear up any terminology confusion, a brief explanation of a few terms might be helpful. First, lets start with the spot price or cash price. These terms are interchangeable and represent the price of a commodity today. This is in contrast to futures or forward prices which reflect the price of a commodity for delivery at a future or forward date. In an arbitrage free world, there is only one spot or cash price, but there can be as many future/forward prices as there are days to deliver in the future. Forwards are simply over the counter (non-exchange) futures and while they are different, for purposes of this discussion we will use them interchangeably. Exchanges set standard dates for contract expiration and delivery periods and since most commodity ETFs hold exchange futures, we will focus on futures contracts.

The shape of a commodity curve is composed of a spot price and all the future prices at each available deliverable period. There are generally two schools of thought as to what futures prices reflect. The first is computational and suggests the futures price must equal the spot price plus some “cost to carry” the physical asset in an arbitrage free environment. The second says that the futures price is the expected spot price at the delivery date in the future. Regardless of definition, what is most important to the long holder, is the shape or term structure of the forward curve.

When the shape of the curve is downward sloping from the spot price, the market is said to be in normal backwardation or backwardation. When it is upward sloping from the spot price, it is said to be in normal contango or contango. The distinction between normal backwardation and backwardation and normal contango and contango is subtle and beyond the scope of this note. We will use the terms backwardation and contango generically to describe the the shape of the curve.

Chart 1

Looking at the chart above, if the market is in a state of contango, futures contracts are above the spot price. If the spot price of the commodity does not move (100 in this example) the futures price will converge on the cash price as time passes. This convergence lower in price is known as negative carry. The opposite can be said when a market is in backwardation. In backwardation the futures prices converges higher in price as it approaches spot and receives positive carry. In other words, if the market is in contango, the spot price must rise to the price of the future for the holder of a long futures contract to break-even.

Compounding the problem is the fact that often, the slope of the curve is greatest nearer to spot or today.  As illustrated, the speed and magnitude of convergence accelerates the nearer you become to expiration or the spot price. This is analogous to being long a call or put option, where the time premium (theta) embedded in an options value erodes quicker as the option nears expiration. For ETFs that only hold futures nearest to spot, this can be particularly troublesome. These ETFs are consistently rolling into contracts at the point on the forward curve that is converging the quickest. This can be a real benefit when markets are in backwardation, but unfortunately markets are spending more time in contango and at an increasing pace.

Measuring Backwardation & Contango

Defining the forward curve is pretty straight forward, however,  measuring backwardation and contango is not quite so easy. Unfortunately, spot or cash data is rarely marked at the same time as an established futures closing price. To insure consistency and accuracy, we measure the shape of the forward curve by observing only futures prices. We use the nearest or front month contract as a proxy for the spot price and measure the shape of the curve using the deferred contracts following the front month. Sixteen individual commodities were observed from five different sectors. We then calculated three different versions based on time and weighting for comparison purposes. Additionally, each chart is sector weighted to reflect the general composition of most legacy commodity indices such as the GSCI, BCOM, and RICI. Sector weighting is equal to; Energy 43%, Grains 19%, Softs 7%, Livestock 7%, and Metals 24%. The charts below describe the percentage of backwardation or contango of the deferred contacts as compared to the front contract.

  • Chart 2 – Front versus next nearest only
  • Chart 3 – Weighted average of a 12 month futures period (front weighted)
  • Chart 4 – Simple average of 12 month futures period
Chart 2
Chart 3
Chart 4

As can be observed in the three charts, since 1997 commodities have generally been in a state of contango. It is also appears (particularly in the 12 month charts) that the percentage of time in contango versus backwardation is increasing. We can also observe that the absolute levels of contango appear to be greater than the absolute levels of backwardation. Table 1 below compiles our observations broken down by full period, and the first and second half of the data:

Table 1


What is most striking, on average the percentage of time commodity markets have spent in contango has increased from 62.4% to 84.7% between the two time frames. Secondly,  while the average level of backwardation versus contango does appear to be declining slightly, the absolute level of contango has consistently been about twice the level of backwardation. Looking at the slight decline in the absolute levels, we suspect this is a result of the decline in overall commodity prices seen over the period.

Perhaps what makes these elevated periods of contango more of a puzzle, it has occurred as interest rates have declined following the credit crisis. A key component to forward prices is the financing or interest rate portion of carry cost. As interest rates rise, other things held steady, carry cost increases theoretically raising the value of forward commodities. Rising interest rates don’t occur in a vacuum and they impact other commodity pricing factors such as currency rates and anticipated inflation, but given interest rates in the U.S. are expected to rise, this could further compound the contango problem.

Another observation to be made from our research, is the degree of backwardation and contango varies considerably among commodity sectors. In terms of curve volatility and magnitude, energy and livestock are the most volatile while metals exhibit the least amount of curve volatility.  The chart below shows backwardation/contango broken out by commodity sector using the weighted 12 month data.

Chart 4

Of note is the volatility and degree of contango in the energy sector over the data period. Except for livestock, energy exhibits the largest swings from backwardation to contango. During the first half of the data, energy was in contango roughly 48% of the time, during the second half it climbed to 81%. This 69% increase in time in contango is material for many broad based ETFs as they are production/consumption weighted making them “energy heavy”. It is even more worth noting for single energy commodity ETFs. To be certain, spot energy prices have collapsed from their 2008 peak and the recent positive term structure may in fact only be a temporary reflection of new lower base prices. However, baring a significant rise in energy prices this may persist for some time.

Contango & Returns

Explicitly tying the shape of the forward curve to returns is difficult at best but evidence suggests that over long periods of time, persistent contango will diminish long position returns. The Pearson correlation between a positive forward curve and commodity returns is only slightly negative ranging from -.15 to -.27 depending on curve method observed. Regression analysis is less informative as while it appears that the shape of the curve might explain some portion of return, r-squared values are near zero. Ultimately, outright market noise drowns out much of the subtle effects of forward term structure to draw any conclusions from these tests alone. Measuring the performance of particular roll methods however offers more insight.

Two better gauge the impact contango may have on a portfolio of commodity contracts, we constructed two baskets of long only commodity futures. Both baskets contain the same sixteen commodity contracts, are equally weighted and rebalance at the same time. Basket one however holds only the front month future and rolls to the next nearest contract (static roll). Basket two employs the ABSR Dynamic roll methodology and holds the futures contact that has the least amount of “negative carry yield” or most amount of positive carry yield on the forward curve (dynamic roll).  Chart 5 below plots the growth of $1,000 for both the static and dynamic roll methods. Directly below chart 5 is the weighted average 12M graph of backwardation / contango (chart 3 above) for visual comparison.

Chart 5

Chart 5 shows a significant period of divergence that occurs following the period from 1997 through roughly 2004 where backwardation and contango existed at a more even rate. From 2005 forward however, the static roll method significantly under performed its dynamic roll competitor. Breaking down the returns into buckets of months in backwardation and months in contango, we find that the average return difference (Dynamic – Static) during periods of backwardation was only 0.015%. This is what we would expect to observe as during periods of backwardation the dynamic roll method would most likely hold the same front month contracts as the static method and therefore have very similar returns. The average monthly return differential between the dynamic and static rolls during periods of contango, however, was 0.375%. In other words by holding contract months further out the curve during periods of contango, the dynamic roll was able mitigate a significant portion of negative carry.


What we can conclude from our construction of curve structure data, is commodity markets are spending a considerable portion of time in contango. It would also appear that since 1997 the amount of time has increased on average by roughly 35% and by more than 69% in the energy sector. We have also shown that markets in a state of contango have a material negative impact on long only commodity portfolios returns. Investors employing commodities as a diversifying asset class to a traditional portfolio should evaluate the roll method their ETF/ETN employs. If this contango conundrum continues to persist at the elevated levels we have seen for the past 10 years, a dynamic roll method is more likely to protect against the unwanted costs of negative carry.

Sources: Charts and Tables: ABSR Research, Futures Data: Commodity Systems Inc.

Shawn Bingham is a 25+ year veteran of the futures and options industry. He was the co-founder of Midwest Trading Partners LLC, a boutique managed futures business catering to HNW clients and fund of funds. Mr. Bingham has recently launched abs(R) Research, a niche firm dedicated to designing next generation alternative asset indices and benchmarks for the investment management community.

©2017 ABSR Research. All rights reserved. This document is intended for informational purposes only, it does not constitute an offer or solicitation to buy or sell any security.  While we make every effort to ensure the accuracy of the information, no guarantees can be made that all information is accurate and complete. This report and corresponding data are for use with institutional or qualified eligible persons only, and not for use of non-qualified investors.
Tagged , , , , , .