Curve Shape vs Price Trend
Author: Brynne Kelly (w/Lee Taylor technical levels)
There are two different ways one can look at calendar strips: either as an outright price average for the 12-month strip or as a 'curve' with a shape that contains information about how market participants view the future. The chart on the left below represents the daily settle price for the 12-month calendar 2021 strip and the chart on the right represents the calendar 2021 'curve' settlements by month for Jan-21 - Dec-21 (yellow = 1/02 settlements, teal = 5/01 settlements and purple = 10/16 settlements).
Obviously we can see that as the curve moved from backwardation to contango (yellow line vs teal line, above right), the average price for the calendar 2021 strip fell significantly. With such a wide range in price levels this year, the details of the curve structure get muted. Since we know the events that led to the huge drop in prices and curve structure, we omit the first 4 months of curve settlements in order to get a closer look at how the curve structure and prices have recovered since the initial drop.
In all 3 cases, the teal line (which represents 7/01 settlement prices) is the flattest structure. The inability of the calendar strips to hold on to that flat of a structure caused prices to begin to rollover. 10/16 is compared to previous curve settles since 5/01. Comparing Friday's slope to that on 5/01, we see that that the curve has indeed flattened out (the slope of the curve on 5/01 is much steeper than on 10/16). The curve reached its flattest structure in the above chart on July 1. The flattest curve above is achieved on 7/1 (line 3 vs teal line above). After July, the curve once again began to steepen (more contango, lines 4 and 5 above, 8/03 and 9/01 settlements, respectively). As the the contango of the curve lessened, the average settlement price for calendar 2021 continued to rise (left chart, above). As the structure was unable to hold on to the flatter July structure, the average price of the calendar 2021 strip began to rollover (left chart, above).
We see the same pattern in both gasoline and distillate futures for calendar 2021 below.
In all 3 cases, the teal line (which represents 7/01 settlement prices) is the flattest structure. The inability of the calendar strips to hold on to July's curve flattening led to a selloff in calendar 2021 prices across the complex with a slight rebound in October
This rebound is best viewed through the lens of 1-month calendar spreads. The level of these spreads represent the expectation of storage economics. We experienced extreme contango conditions earlier this year when demand dropped and barrels rushed to find a home. With that in mind we look at 1-month calendar spreads. The chart on the left shows the trend of 1-month calendar spreads since June (month-1 vs month-2, month-2 vs month-3 and so on) and the chart on the right shows the shape of 1-month calendar spreads for the next 24 months.
At the beginning of June and July (gold and aqua lines above right) monthly spreads were fairly similar across the 24-month strip. By August, however the front spreads had weakened much more than the back and the shape of the spread curve got much steeper.
What's interesting to note is that the front spread (month-1 vs month-2) continues to show relative strength. Contango that is 'less weak' in the front spread is not, however, providing much overall price strength, let alone any direction for the rest of the curve. Either way you look at it, these spreads are relatively small (ranging from -$0.38 to -$0.03). This suggests an expectation that there will be plenty of traditional storage space available (contango levels that we saw earlier this year were a reflection that traditional storage space was not available and the market needed to price in alternative more expensive storage space).
Why does the market think traditional storage space will be available?
The Impact of Delta on Gulf Coast Inventories and Refinery Utilization
Last week's inventory report was a confusing one for sure. For the week, total inventories were down 13.70 million barrels. Of that total, commercial crude oil inventories were down 3.70 million barrels for the week. The majority of the draw-down came from products, which drew 8.80 million barrels on the week. The remaining 1.20 million barrel draw came from the Strategic Petroleum Reserve (see Weekly Inventory section for detail). This draw in crude oil inventories happened during a week when refinery utilization was down 2%, leading to less crude oil being input to refiners for processing (yellow line below = 2020 input to refiners).
Lower refinery utilization coupled with draws in both gasoline and distillate inventories suggests that production is not enough to meet current product demand. We have seen this in previous weekly reports in the gasoline numbers, however the majority of the draw was in distillate inventories. If we could continue to operate refining capacity at this lower rate, we might be able to make significant headwind in drawing down product inventories.
So, why did crude oil draw during a week of lower refiner utilization? We can thank production shut-in's from hurricane Delta for that. For the week-ending 10/9, there were a total of 5.42 million barrels of production offline in the US Gulf of Mexico. PADD 3 inventories (US Gulf of Mexico region) were down 5.20 million barrels for the week.
If you add this to last week's draw of 3.7, you get a net build of 1.72 million barrels for the week. Cushing posted a build of 2.90 million barrels for week ending 10/09. For next week's report (week ending 10/16) we know that there was another 6.62 million barrels offline during that period per the BSEE. If refiners hold at the same utilization rate reported for 10/9 (75.10%), we would expect to see draws across the board in the next report.
Despite this, WTI could not hold on to a rally and prices pulled back by the end of the week (gold line below).
The same was true for both gasoline and distillate markets (again, yellow line in both charts).
Given that we are moving towards year end, we have a historical trend to deal with. This is due to ad-valorem taxes. According to the EIA:
"At the end of December each year, parts of Texas and Louisiana, where significant volumes of crude oil are stored, assess ad valorem taxes (meaning, according to value) on end-of-year crude oil inventories. These taxes, along with the generally accepted accounting practice of last-in, first-out (LIFO) method used to value the assets, create an incentive to draw down crude stocks in the region at the end of the year in order to reduce the tax bill. If oil prices have risen during the year, this accounting practice gives companies stronger incentive to reduce inventory because doing so will further limit their tax exposure. Conversely, if oil prices have fallen throughout the year, companies have less incentive to reduce crude held in storage."
We can see that PADD 3 inventories have fallen significantly from their highs in May.
We expect that these will fall further in next week's EIA inventory report given the BSEE stats noted earlier. In addition, any increase in refinery utilization - absent an increase in US production - will lead to inventories falling back into their 5-year range. Typically the selling of inventory towards year-end leads to weaker calendar spreads. However, we think this year could be different and front month spreads will remain relatively strong. This will be a result of hurricane Delta-induced inventory reductions and 2020 hurricane season-induced refinery outages that helped to reduce product inventories.
EIA Inventory Statistics
The EIA reported a total petroleum inventory DRAW of 13.70_million barrels for the week ending October 9, 2020 (compared to a draw of 3.20 million barrels last week). Of that total, commercial crude oil inventories were down 3.70 million barrels for the week.
Year-to-date, total inventory additions stand at a BUILD of 63.00 million barrels (vs 79.60 last week).
Commercial Inventory levels of Crude Oil (ex-SPR) and Distillate are still above their 5-year average,
Lee Taylor - Technical Levels
(To return next week....)