Crude Oil Backwardation: Theory, Facts And Myths

This article is a follow-up to one I wrote a couple of years ago, "When Do Crude Futures Flip Into Backwardation?" I updated the statistics and added a section on price implications. I have noted in my reading of articles and in readers' comments a strong belief that backwardation leads to price increases, but my analysis disputes that theory.

For non-futures traders, contango is a futures price market structure in which future prices are higher than the nearby contract. A backwardated market is one in which future prices are lower than the nearby contract.

https://static.seekingalpha.com/uploads/2017/2/13816732_14862488407130_rId5.png

"Normal Backwardation"

John Maynard Keynes was one of the most influential economists of the 20th century. He developed theories on the causes of business cycles and advised on the economic policies of governments.

In Treatise on Money (1930, chapter 29), Lord Keynes argued that in commodity markets, backwardation is a normal market situation, and so he referred to it as "normal backwardation." Producers of commodities are more prone to hedge their price risk than consumers, and so there is more selling than buying interest after the nearby month.

All things being equal, this also makes sense from an insurance risk standpoint. Producers want greater certainty for their future revenues and are willing to pay a risk premium. Speculative buyers are willing to take the risk, but they want to be paid a risk premium to do so.

I initially reviewed WTI crude oil futures weekly time spreads (Month 1-Month 4) for the 10-year period from January 2007 through January 2017. I found that the market was in contango 76 percent of the time.

https://static.seekingalpha.com/uploads/2017/2/4/13816732-14862490920942414_origin.png

U.S. crude oil stock levels are one important measure of whether the market is undersupplied or not. An undersupplied market is another factor that could create backwardation, in addition to the risk premium Keynes wrote about. Low stocks increase the potential for a supply disruption, causing oil prices to spike.

I constructed a frequency distribution to assess the conditional frequency of each market structure depending on the stock levels. For example, the "Count" is 100% for 265 million barrels and above. Backwardation was present 24% and Contango 76%. At the other end of the table, stocks were 500 million or higher in only 2% of the weeks, and prices were in contango 2%.

January 2007-January 2017

Historical Frequency Distribution

Inventories (mmb)

265

300

350

360

393

400

450

500

Count

100%

90%

32%

24%

19%

19%

13%

2%

Backwardation

24%

18%

4%

2%

0%

0%

0%

0%

Contango

76%

72%

28%

22%

19%

19%

13%

2%

Note: The statistics refer to the percentage of time that inventories were at that level or higher. The midpoint was 393 mmb.

Source: Boslego Risk Services

Based on this analysis, with crude stocks at 393 million, its five-year average, prices had been backwardated 0% of the time in that condition. Inventories had to be about 360 million barrels or less for backwardation to be present.

A third reason backwardation may exist is a "disruption risk" premium. By this, I mean the uncertainty related to potential disruptions of oil supply, unrelated to current inventory levels. Speculators demand a premium to go short the nearby contract under that condition.

This last factor may greatly complicate traders' interpretations of market price expectations. It could be present when current inventories are not undersupplied, but when the market demands a "disruption risk" premium.

In fact, this WTI backwardation was present in two recent weeks (discussed below) when crude stocks were more than adequate. I believe that uncertainties about production and exports from both Iran and Venezuela had created the "disruption risk" premium.

January 2009-April 2019

I updated my analysis to cover a more recent 10-year period as stated above. I computed a new frequency distribution as shown below:

Inventories (mmb)

265

300

350

360

393

400

450

500

Count

100%

100%

53%

45%

41%

40%

23%

6%

Backwardation

23%

23%

13%

11%

9%

8%

0%

0%

Contango

77%

77%

40%

35%

32%

32%

22%

6%

Note: The statistics refer to the percentage of time that inventories were at that level or higher.

Source: Boslego Risk Services

There was a shift higher in the frequency of backwardation as a function of stocks; i.e., more backwardation was present at higher stock levels. This is logical in that higher stocks are required to provide the same level of demand cover, as measured by refinery inputs of crude oil.

However, backwardation was present in two recent weeks (April 12th and April 26th) while stocks were high. In a true-false test, if stocks were above 450 million and backwardation was present, those were the only weeks since 2009:

Time Spread Test

Source: Boslego Risk Services

This was the period just ahead of Trump's announced decision to terminate the waivers to Iran sanctions. Clearly, the market was concerned about a potential disruption to supply and required a "disruption risk" premium.

Price Implications

I frequently read in articles and in reader comments the belief that the presence of backwardation implies oil prices are likely to rise. I tested that theory empirically and found that is, in fact, not the case.

I performed a linear regression of the time spreads to subsequent price changes over the following 1, 2, 3 and 4 weeks. I found that the relationship is actually slightly the opposite; i.e., that prices were more likely to drop! However, the r-squared values are so low that neither a price gain nor a price drop should be expected. The T-ratios of the coefficient were also statistically insignificant.

# Weeks

Coefficient

R^Squared

1

-0.06828

4%

2

-0.07286

2%

3

-0.07885

2%

4

-0.07979

2%

Conclusions

There is evidence that backwardation is more frequent when crude oil stocks are at lower levels. However, there is no evidence that oil prices are more likely to rise over the next one to four weeks, if backwardation is present.

Therefore, it is a myth that time spreads lead price changes because they may simply reflect a risk premium, as theorized by Keynes or reflect a "disruption risk" condition I explained above.

My theory is that time spreads are a result of market risk premiums, not a leading indicator of future prices. The market is more sophisticated for that to be true; i.e., it would be too easy to make money trading if the price change could be predicted from the time spreads, and so that effect is arbitraged out of existence.

The empirical data analysis supports my view that time spreads do not provide predictions of future price changes. Therefore, I do not view them as a determinate in trading.

A case in point: Brent crude was/is in backwardation:

Source: Barchart

But the price was crushed Wednesday and Thursday:

Source: Barchart

Using time spreads to predict future oil price changes is no better than consulting a fortune teller using tea leaves.

Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Need more info about the Publisher?

This article was generated with the help of our AI and we are simply sharing the news. Click here to find out more information about this article & its publisher. CreateAI takes no responsibility for the accuracy or use of any or all of this content.

Want to enhance your online marketing?
We can help!

Automated Content Marketing!
Social Media Marketing & SEO