A Refined Analysis of Crude Oil

Below, I wish to provide an analysis on WTI crude oil that is hopefully fresh, original, and most importantly thought provoking.

Let’s start with what we know: positioning data

Fortunately, the Commodity Futures Trading Commission provides us with a detailed weekly summary on positioning data through its Commitments of Traders (“COT”) report and based on that report it is a well known fact, at least among sophisticated traders, that the net positioning of speculators is, and has been for a while, at record levels going back more than a decade. Chart below:

wti-crude-oil-speculative-positioning
Click on the chart above for a bigger chart

A first level analysis of this chart is obvious — the historically non-sticky (read trigger happy) side of the market is extremely long and will be looking to rush to the exits at the first whiff of this market turning down, causing a stampede on their way out. There is certainly some truth to this but lets take our analysis a little deeper. We’ll come back to it.

Let’s focus on price for a second

crude-price
Click on the chart above for a bigger chart

A quick look at the price chart (below) suggests that the market is forming a bottom. All signs point to that — 1) a bounce from $25; the price point from which the market started its longest and biggest rally, in early 2000, 2) an inverted head and shoulders bottom, with its head at $25 and its neck line at around $40.

“Markets tend to return to the mean over time.”

– Bob Farrell’s rule # 1

Now, lets refocus on the price chart but with some fancy modifications. Before we go ahead it would be grossly wrong of me to not take a moment to tip my hat to those who deserve it:

ht 1) to Michael J. Oliver, a product of E.F. Hutton & Co., for teaching us this wisdom, and 2) to RealVision TV (https://www.realvision.com/), for bringing to us people like Michael, and for all the other fantastic work they put out.

crude_msa
Click on the chart above for a bigger chart

The 1st panel in the above chart (in blue) is a price chart, of WTI crude oil, alongside its long term moving average and the 2nd panel (in green) is a chart that plots the price’s distance from its rolling mean. As we can see from the 2nd panel in the above chart, deviations from mean tend to exhibit characteristics of not only being range bound but also that of trending up/down. This not only affirms Bob Farrell’s Rule # 1 but also provides us with an additional tool to analyze the price action.

The price of WTI crude oil, 1st future, is around 13% below its long term mean (at its worst level it was about 67% below its mean) but, as you can see in the chart above, it is bumping into resistance from a downtrend line. To elaborate on this, the deviation of price from its mean has been in a steady decline since 2011. In 2011, at its highest point, the price was 43% above its mean; in 2012, at its highest point, it was 32% above its mean; in 2013, at its highest point, it was 19% above its mean; in 2014, at its highest point, it was 12% above its mean; in 2015, at its lowest point, it was 57% below its mean; in 2016, at its lowest point, it was 66% below its mean. The downtrend line on this chart is around -13%.

So, unless the ‘deviation from mean’ of WTI crude oil is able to overcome this resistance, it is hard to agree with the price-only analysis and conclude that price is indeed bottoming. This could very well be a bear-market rally that is misleading us into believing that a bottom is in place. As Jamie Coleman (founder of Forexlive) used to say: “If you try to catch bottoms, all you’ll be left with are smelly fingers.”

Let’s focus on fundamentals for a second

The EIA recently published its short term energy outlook (“STEO”) report, some highlights below:

Global Petroleum and Other Liquids

  2015 2016 2017 2018
Supply & Consumption

(million barrels per day)

Non-OPEC Production 58.8 58.2 58.5 59.6
OPEC Production 38 39 39.6 40.2
OPEC Crude Oil Portion 31.6 32.5 32.7 33.2
Total World Production 96.8 97.2 98 99.8
OECD Commercial Inventory (end-of-year) 2969 3087 3066 3107
Total OPEC surplus crude oil production capacity 1.46 1.26 1.69 1.21
OECD Consumption 46.3 46.6 47 47.3
Non-OECD Consumption 48.7 49.8 51.1 52.3
Total World Consumption 95.1 96.5 98.1 99.6

According to EIA’s latest STEO, the oil market seems to be well balanced at these levels and whatever increases they have forecasted in consumption (from their previous STEO) seems to be stemming from an increased baseline effect due to revisions to their Chinese & African consumption data for 2013-2014. Furthermore, their projections are based on the following two assumptions (debatable):

Primary Assumptions (percent change from prior year)
  2015 2016 2017 2018
World Real Gross Domestic Producta 2.6 2.3 2.7 3
Real U.S. Dollar Exchange Rateb 10.7 6.3 5.3 0.9
a Weighted by oil consumption.
b Foreign currency per U.S. dollar.

The recent OPEC agreement was structured in such a way that members agreed to the cuts on a six-month basis, which means every six months these guys have to get together and agree on it, I’d place very low odds of this agreement lasting for any meaningful length of time.

Add to this any surprise in US production growth and the oil market could easily be imbalanced. Below is a chart that plots US crude oil production. At its highs, US produced 9.6 million barrels/day around June-2015. As I’m sure many of you are aware this was the time around which rig counts started to drop rapidly and production dropped by 12.5% to 8.4 million barrels/day around June-2016. It has currently risen from its lows by 6.9% to 8.98 million barrels/day. If you have the slightest bit of confidence (I have plenty) that US shale producers are constantly improving their technology then you’d have to conclude that they are continually lowering their marginal cost of production, which means they can continue to remain profitable at much lower oil prices, which in turn means that they will be looking to capture more and more of the total market share; look for any meaningful rally in crude oil to be met with substantial supply from these US cowboys.

us-crude-oil-field-production
Click on the chart above for a bigger chart

One more fundamental chart to consider: relationship between the price of crude oil and crude oil inventories

crude-inventory-data
Click on the chart above for a bigger chart

The 1st panel in the chart above plots the rate of change of inventories (less SPR) and the 2nd panel plots the rate of change of WTI crude oil. You can see that going back to early 2000, there are three distinct time frames in the 2nd panel where the rate of change of crude oil was persistently negative: 1) May-2001 to June-2002, 2) Oct-2008 to Oct-2009, 3) Jul-2014 to Aug-2016.

Every one of these time frames was followed by an explosive rally in crude oil, where the rate of change of crude oil remained persistently positive. The difference is that the first two time frames coincided with US recessions and following both these time frames the rates of change of inventories entered into negative territory but the third time frame did not coincide with a US recession (this was purely a supply shock) and nor was it followed by a period where the rate of change of inventories went negative (no demand surprise either). Another flatulent whiff of a bear market rally.

Circling back to positioning data; lets dig a little deeper

The first thing we need to recognize is that while the total Open Interest (2nd panel in chart below) in WTI crude oil is at its all time highs, the total nominal value of all open contracts (1st panel in chart below)  in WTI crude oil as measured in USD terms has barely retraced 50% of its 2008 & 2014 levels.

crude-oil-price-vs-oi
Click on the chart above for a bigger chart

ht to Kevin Muir of the Macro Tourist, http://themacrotourist.com/macro/re-evaluating-crude-oil-spec-positioning, for bringing this to my attention. If you aren’t a subscriber to his blog, I highly recommend it.

While the chart above certainly tones down the perceptive froth in the oil market and many would contend that this is a reason to not be overly worried about the net speculative positioning, I contend that the positioning data should be used to analyze crowd behavior NOT froth. The reason to look at positioning data is to identify how crowded the market is and the chart below says it all.

The gap between non commercial longs relative to non commercial shorts, both measured as a percentage of OI, is above its 95th percentile. Hardly any financial participant is not already long.

wti-net-positioning
Click on the chart above for a bigger chart

We dig even deeper and we, finally, find gold 🙂 the ADF indicator

Everything I have presented so far is just analysis of data. What I’m going to show you now is pure logic. Before I begin, let me start by tipping my hat to Albert D. Friedberg (“ADF”), of the Friedberg Mercantile Group, the man who identified this logic. Albert was one of the first board of governors of the Toronto Futures Exchange, and served as its Chairman from March 1985 to June 1988. He currently manages The Friedberg Global Macro Hedge Fund and the Friedberg Asset Allocation Fund. Over the years he has been kind enough to have coloured my understanding of global-macro investing and to this day continues to be my global-macro teacher.

Of all the arguments I have heard against my bearish case in oil, the only one I haven’t been able to successfully rebuttal is ADF’s argument. His argument is simple: “Producers have oversold”.

Let me elaborate: When we (referring to the entire analyst community) analyze the COT data, we always focus on non-commercials and it makes sense to do so considering that it is the speculators that are trigger-happy (they tend not to be sticky with their holdings) but the other side of speculators are commercials and in this case it is mostly producers (it can also be huge consumers like airlines, etc. but we know from the quarterly filings of most airline companies that they’ve stopped hedging their oil needs a while ago, at least in a meaningful way). So, one way to look at the large net-long speculative positioning is that the speculators have over-bought but another way to look at it is that the commercials have oversold. The reason this argument has merit is that the US shale revolution (read supply glut) is a well known fact, NOT news. Ask yourself, if you were in the business of producing a commodity that is in abundant over-supply, how do you protect your downside? Simple, you hedge; not just what you have produced but also what you expect to produce over a reasonable horizon.

A million dollar question: how do you know if speculators have over-bought or if commercials have over-sold?

This is where the genius of ADF comes into play; he came up with a simple logic:

“Measure the rate of change of Open Interest (a measure of total barrels floating around in financial futures) relative to the rate of change of inventory (a measure of supply). If the gap is increasing (more barrels of oil is finding its way into financial futures than into inventory) then the market is oversold (by producers) and should be well bid (by speculators) but if the gap is shrinking (more barrels of oil finding its way into inventory than into financial futures) then the market is overbought (by speculators) and should be well offered (by producers).”

– ADF

Below is a chart that captures this simple logic. A picture is worth a thousand words:

oi-vs-inventory-adf-indicator
Click on the chart above for a bigger chart

The ADF indicator is plotted on the 1st panel and the price of WTI crude oil in the 2nd panel. As you can see in the chart above, the ADF indicator works brilliantly in identifying major turning points going back over a decade. The indicator turned around lower in late-2007 and oil peaked in mid-2008, it turned around higher in early-2009, which coincided with a turn around higher in price, it turned around lower in mid-2013 and the oil market fell off a cliff in June-2014, it turned around higher in early 2016, which coincided with a turn around higher in price.

Considering that the key to using this indicator is to measure turning points in the indicator itself, I have plotted multiple (12 months, 6 mo. 3 mo, and 1 mo.) rates of change of the ADF indicator in the chart below.

adf-indicator-rate-of-change
Click on the chart above for a bigger chart

The ADF indicator  seems to be rolling over as well. As of this writing, the 1 mo. rate of change has already turned negative: -9.64%. The 3 mo. (+4.9%), 6 mo. (24.8%), and 12 mo. (21.6%) rates of changes are still in positive territories albeit declining rapidly. The ADF indicator could very well be screaming a turning point in oil. The fact that consensus is extremely bullish, price has barely reared its head from a 15 year low and the ADF indicator is already rolling over confirms to me that this rally in crude oil is nothing but a vicious bear market rally that has sucked in every last financial participant around.

Coming back to the price chart

As a student of the markets the one thing that I can say with great confidence is that we should always respect the price chart.

crude-price
Click on the chart above for a bigger chart

It is clear from the price action, since the lows in early 2016, that there have been a string of higher highs and higher lows; this is a bullish development. If this action continues then the rising trend line, connecting those highs, will most likely bump into strong resistance at 61.4. This would be the time to pull the trigger to go short crude oil. The risk is that a war breaks out (or some unknown, unknown, happens) and it busts through a higher resistance level at 77.1 and so, this level would be the obvious stop loss. The objective is to capture a retracement to the lows of early 2016 at 25.9.

Happy trading!