Crude Oil: Don't Get Suckered Into Buying This Dip, Oil Has Further To Fall
Oil rose for the first time in 5 days yesterday after the EIA reported a larger-than-expected drop in crude oil inventories and the largest slide in domestic production since July.
However, I argue in this article that the factors that led to these surprising declines are ephemeral and that the underlying supply/demand picture remains bearish.
I believe that a pullback to $50-55/barrel is not only inevitable but will be beneficial for the long-term health of the burgeoning crude oil rally.
My trading strategy is discussed in detail.
After surging by 30% to over $60/barrel since bottoming at $43/barrel in mid-March, crude oil has spent the first three weeks of May in choppy range-bound trading as it awaits a catalyst to justify a further move higher. On Tuesday, thanks to a surging US dollar and concern over global oversupply, crude oil tumbled over 3% in the largest daily decline since April 8th to $57.26/barrel, a 6% pullback from recent highs and the lowest closing price for WTI crude since April 28. It marked the fifth consecutive day of losses and left bulls scrambling to defend the sustainability of the rally.
The EIA appeared to come to their rescue yesterday as a larger-than-expected inventory decline in crude stocks and the largest slide in US production this year snapped the losing streak with oil rallying 1.7% to $58.98/barrel and gasoline rising 2.3% to $2.04/gallon. The popular ETFs, the United States Oil Fund (USO-NY) and the United States Gas Fund (UGA-NY) rose in turn 1.1% and 1.3%, underperforming thanks to a late afternoon pullback after the commodities markets had closed. Nonetheless, while this week's inventory report again appeared to show bullish data across the board, reading between the lines paints a much more bearish picture that is currently being masked by isolated short-term events. This article analyzes yesterday's EIA Petroleum Report and discusses the short and long-term prospects for the commodity.
In its weekly report, the EIA announced that 2.7 million barrels of oil were withdrawn from storage during the week ending Friday, May 15. It was the largest weekly withdrawal since the week ending November 28, 2014. Total storage now sits at 482 million barrels. Since peaking at 491 million barrels the week ending April 24, storage has now declined for three consecutive weeks. The 2.7 million barrel storage draw was well above the five-year average 1.2 million barrel draw, but was below last year's mammoth 7.2 million barrel draw, which was the second largest of the entire year. The storage withdrawal was also greater than the 1.1 million barrel draw projected by analysts, but noticeably below the American Petroleum Institute's (API) Tuesday announcement of an expected 5.2 million barrel draw. Nonetheless, the current storage surplus versus the five-year average declined by 1.5 million barrels and now sits at 102 million barrels, which is the smallest surplus since the week ending March 20. Figure 1 below shows current storage levels versus 2014 levels and the 5-year mean highlighting the ongoing storage surplus.
(click to enlarge) 
Figure 1: Crude oil inventories versus the 5-year average and 2014 showing a slow decline in the storage surplus over the last 3 weeks.[Source: EIA Petroleum Weekly]
Unfortunately, a closer examination of the storage withdrawal paints a less rosy picture.
The US is divided into 5 PAD - or Petroleum Administration For Defense - Districts. District 1 is the Atlantic Seaboard, District 2 the Great Lakes, Midwest, and Northern Plains, District 3 Texas and the Gulf Coast, District 4 the Rockies, and District 5 the Pacific Coast. Figure 2 shows a map of these regions.
Figure 2: Maps of PAD Districts [Source: EIA]
The majority of US oil is produced in PAD Districts 2 and 3 and storage levels in these regions give the best picture of the US supply/demand picture while PAD Districts 4 and 5 tend to be isolated from the rest of the country thanks to poor interstate pipeline networks across the Rockies and are more susceptible to external determinants of supply and demand such as imports.
Figure 3 below shows a plot of last week's crude oil inventory changes by PAD district.
(click to enlarge) 
Figure 3: Crude oil inventory change by PAD District showing flat inventories across PADs 1-3 with steep declines in PADs 4-5 contributing to the national decline. [Source: EIA Petroleum Weekly]
Based on the data above, the entirety of last week's 1.5 million barrel deficit versus the five-year average can be attributed to drops across PADs 4 and 5. While PAD 3 saw a 1.2 million barrel drop, thanks to the extensive pipeline network across the region, these inventories were likely distributed to PADs 1 and 2, which saw a net 1.8 million barrel gain. On the other hand, PADs 4 and 5 saw a net -3.3 million barrel decline in inventories, which more than accounts for the nationwide -1.5 million barrel deficit to the 5-year average weekly withdrawal.
What accounts for this West Coast decline?
One of the big numbers coming out of yesterday's report was that domestic crude oil production slid last week by an enormous 112,000 barrels per day from 9.374 to 9.262 million barrels per day. This was by far the largest drop since domestic production peaked the week ending March 20 and the largest overall since July of 2014. However, on closer examination, that entire decline came from the state of Alaska, which saw production slide from 504,000 barrels per day to 392,000 barrels per day. This represents an absurd 22% decline in statewide production in a single week and likely will be a one-week anomaly, which is ironic given that Shell just this week announced it was resuming its controversial drilling off Alaska's North Slope.
On the other hand, lower 48 crude oil production remained unchanged at 8.87 million barrels per day. Since peaking on March 20, lower 48 production has slid a mere 41,000 barrels per day, or 0.46% despite a 60% slide in the domestic rig count. Figure 4 below shows lower 48 production versus Alaska production over the past year.
(click to enlarge) 
Figure 4: Lower 48 oil production (left axis) vs. Alaska oil production (right axis) showing stabilization of lower 48 production since March peak with steep 1-week declines in Alaska production. [Source: EIA Petroleum Weekly]
Why is this relevant?
Unsurprisingly, much of Alaska's oil winds up on the West Coast via Valdez Bay. If we assume that 100% of Alaskan crude winds up in PAD 4 and 5 and that the 1-year average Alaska oil production is 500,000 per day, last week's precipitous drop would account for 740,000 barrels of lost supply versus a typical week, or about half of last week's 1.5 million barrel national deficit versus the 5-year average.
What about the remaining 750,000 barrels?
Despite a glut of foreign oil, US imports have declined in recent weeks thanks to delayed tanker departures from the Middle East. US imports increased overall last week, increasing from 6.9 million barrels per day to 7.2 million barrels per day. However, this is still 161,000 barrels per day below the 1-year average. Figure 5 below shows US daily imports over the last year versus the 1-year average.
(click to enlarge) 
Figure 5: Crude oil daily imports showing last week's imports remaining 161,000 barrels per day below the 1-year average. [Source: EIA Petroleum Weekly]
Figure 6 below breaks down this 161,000 barrel daily import deficit by PAD district.
(click to enlarge) 
Figure 6: Breakdown of last week's 161,000 barrel per day import deficit by PAD district showing PAD 5 (West Coast) shortfall. [Source: EIA Petroleum Weekly]
Again, while PADs 1-3 see net imports well above the 1-year average, PADs 4 and 5 see a net 260,000 barrel per day deficit versus the 1-year average. This national 161,000 barrel per day import deficit - attributable to significant shortfalls on the West Coast - leads to a 1.1 million barrel weekly supply shortfall, more than accounting for the remaining 750,000 barrel weekly deficit versus the five-year average. In short, blame it on the West Coast. Of note, it remains to be seen how the up to 100,000 barrel oil spill from the Plains All American Pipeline along California's Refugio beach will affect the PAD 5 supply/demand picture in coming weeks, if at all.
In summary, I attribute the unexpectedly large 2.7 million barrel EIA-reported storage decline this week to a decline in imports and domestic supplies on the West Coast. While imports to PADs 1-3 were above average and lower 48 production was again unchanged, imports to PADs 4-5 slumped by 270,000 barrels below the 1-year average and production originating in Alaska slid by 112,000 barrels per day, which together more than wiped out the 1.5 million barrel deficit versus the 1.2 million barrel withdrawal 5-year average. Unfortunately, for natural gas bulls, with Middle Eastern Producers boosting production capacity, imports to the US are likely to trend higher, particularly if crude prices trade back up to $60. Further, last week's unexpected Alaskan supply drop is likely a short-term anomaly given how stable production in this region has been. Once these short-term anomalies correct themselves, I expect domestic crude oil supply/demand balance will again shift to favor the bears.
Looking longer term, the likely determinant of oil price going forwards will likely be domestic production. As noted, despite a 60% drop in the rig count, lower 48 production has slid a mere 0.4% - or 1.6% for all 50 states if you factor in last week's drop in Alaskan output (which I don't). It has to be disconcerting for the bulls to see such a precipitous drop in the rig count with such a small payoff in production declines. As the saying goes, make hay while the sun shines. Despite the sun shining to the tune of 60%, minimal hay has been made.
And the bulls may be running out of time.
Last week, Baker Hughes announced that oil-directed rigs slid by 8 rigs to 660. While this is the fewest rigs drilling for oil since the week of August 20, 2010, the 8 rig drop was also the smallest since the week ending December 5, 2014. I expect that the oil rig count is in the process of bottoming somewhere around 650. Looking back at the last great rig collapse in 2008-2009, the rig count during that event plummeted by 60% as well before rebounding. This data is shown below in Figure 7.
(click to enlarge) 
Figure 7: Baker Hughes Rig count for 2008-2009 versus 2014-2015 oil crashes showing that the current rig count appears to be bottoming very close to when the rig count bottomed in 2009. [Source: Baker Hughes]
If the rig count in 2015 follows a similar pattern, how long before domestic production is again setting new record highs?
Figure 8 below compares natural gas production to the oil rig count during 2009-2010.
(click to enlarge) 
Figure 8: Crude oil production versus rig count in 2009-2010. Despite precipitous drop in the rig count, production reached new highs just 5-months after the rig count bottomed. [Source: Baker Hughes]
After oil production peaked in April 2009 about 6 weeks prior to the rig count bottoming in early June - very similar to our current situation - oil production declined by about 10% as the rig count bottomed and then began climbing again. By November of 2009 - just 5 months after the rig count bottomed - crude oil again hit a new record high.
While the methods of drilling have changed somewhat since 2010, this is a rather alarming statistic for crude oil bulls. Should the 2015 rig count make a V-shaped recovering after its likely bottom in the next week or two, I expect crude oil production to again approach new highs north of 9.5 million barrels per day by the end of 2015. In order for the nearly 102 million barrel surplus to have a shot at being eroded, the crude oil rig count must remain suppressed for an extended period of time. One of the reasons why production has not dipped further despite the 60% haircut in the rig count is that the several thousand drilled but unfinished wells are being completed and initiating production now that crude oil is hovering around the $60/barrel mark that is typically seen as the average profitability threshold for domestic producers. In order to see a prolonged decline in production, the remainder of these wells need to be pushed into production. Due to steep production declines from horizontally drilled rigs, I expect that once this happens, production will decline rather quickly - provided that the number of rigs does not begin to increase. Under this scenario, domestic production will take much longer to recover given that new production must come from freshly drilled wells, not previously drilled wells that just need to be completed.
In order for the rig count to remain suppressed and avoid a V-shaped recovery, crude oil prices need to remain under $60/barrel and even pull back towards $50/barrel in the short-to-intermediate term to render it unprofitable for producers to drill new wells. Therefore, while it may be counterintuitive, I believe that the best way to ensure the sustainability of a long-term rally is for near-term weakness.
In conclusion, I believe that yesterday's larger-than-expected crude oil withdrawal was attributable to short-term factors driving supply and demand primarily on the West Coast that I expect to correct over the next few weeks. With an over 100 million storage surplus, even a return to a balanced supply and demand picture would spell disaster for the commodity. Further, with the rig count poised to bottom over the next few weeks, there is only a 5-7 month window based on historical patterns for crude oil production declines to eat away at the storage surplus. In order for crude oil to see long-term gains, I believe that a sharp correction in crude oil prices to $50-55/barrel is needed to suppress the rig count. Goldman Sachs tends to agree and on Monday released a report predicting $45/barrel oil by October.
Given this thesis, how am I trading crude oil?
Figure 9 below shows a comparison of crude oil prices and the ETF USO since January. Note that since crude oil bottomed, the ETF has done a very good job of tracking the commodity.
(click to enlarge) 
Figure 9: Crude oil and USO January-present showing bottom, rally, and subsequent pullback in crude oil. [Source: Yahoo Finance]
Since peaking at $60.93/barrel on May 6, crude oil had pulled back 6% prior to yesterday's rally. While not the largest pullback since the bottom back in March, the 10-day period since the high is the longest correction.
When oil eclipsed $60/barrel two weeks ago, I began selling short USO and have now built a short position equal to 15% of my portfolio with a cost basis of $20.50 versus Wednesday's closing price of $19.84. On any follow-through from yesterday's rally, I plan to add to this position until the position equals 25% of my portfolio. Should the commodity somehow exceed $65/barrel, I will begin shorting the 2x leveraged ProShares Ultra Bloomberg Crude Oil ETF (UCO-NY) or VelocityShares 3x Long Crude Oil ETN (UWTI-NY) to gain leveraged short exposure. My price target is $55/share or around a 10% drop from current levels, at which point I will begin closing out my position with a goal to have covered entirely should oil reach $50/barrel, or a 20% profit, plus any underperformance in USO due to contango-associated rollover losses. |
