Last weekend the Wall Street Journal carried two interesting headlines: “Dollar is Pressured from Many Directions” and “U.S. Oil-Price Rally is Longest in Years”.
The direction of oil prices is the subject of endless conjecture, questions, speculation, fantasies, and analysis of bad data. As noted in my previous comment, oil prices started a prolonged decline in August 2014 after the Federal Reserve stopped Quantitative Easing – which means they stopped pumping $80 billion per month of new dollars into the financial system. Oil prices declined for four months before the Saudis announced they would not cut their production rate to support prices; they wanted to maintain their market share rather than continue as the swing producer. After that, the price continued to fall but storage rates did not increase above historical levels until 1Q 2015.
Most reports, analyses, and commentators ascribe the fall in oil prices and their failure to regain previous levels to increased domestic US production resulting from widespread use of improved techniques of horizontal drilling and hydraulic fracturing in very low-permeability rock formations which could not be developed and produced economically with older methods. The logic was that an excess of oil production rates above consumption would cause price decreases in a marginal pricing system on an open exchange (both the NYMEX and the London ICE) and that increases of US production had caused such an excess capacity. This excess capacity became commonly referred to as a “Glut”. The trouble with this logic was that the first indication of excess production was several months after the initial price decline.
Not only that, but The Glut was quite elusive; it was very difficult to find and quantify. Most attempts to determine its size eventually estimated that world excess production capacity was somewhere near 1.5 million barrels per day. Total world production rate was about 95 million barrels per day so the excess capacity was about 1.5% of total production. This seems rather small to be classed as a Glut ( a Mini-Glut?) especially when one considers world production rates, storage volumes, and consumption rates are not known with an accuracy of 1.5%.
For a while, quoted oil prices increased and decreased based on estimates of production rates, decline rates, rig counts, consumption, and other indications of changes of the excess production rate – inherently an exercise in futility because the excess is a small difference between two large numbers, production rate and consumption, which are not known accurately.
Many commentators conjectured the Saudis expected low oil prices suppressed by continuing production at historical rates would put many American oil companies out of business and cause a quick and significant reduction of US production rates. In fact, many American companies entered bankruptcy but US oil production increased through 2015 and then started a slow decline. It seems the Saudis were possibly not aware of the extent of price hedging by American companies or the idiosyncrasies of American Chapter 11 bankruptcies.
By the end of 2016, Saudi Arabia realized that lower oil prices had reduced their national income considerably but had little effect on worldwide production rates which were still in excess of consumption by about 1.5%. The Saudis then organized an agreement to reduce production rates by 1.2 million b/d with several other OPEC countries and 600,000 b/d with Russia and other non-OPEC countries. OPEC members Nigeria and Libya are exempt from the agreement. Compliance with the agreement was better than expected which put some upward pressure on prices but US, Nigerian, and Libyan production rates have increased by amounts offsetting the cuts and prices recently have declined again.
Even oil market traders and journalists eventually realized the difficulties inherent to estimating excess production capacity as an indicator of long-term oil-price trends. The Glut was re-defined as the excess of oil in storage above historical volumes. With this logic, one could determine whether the price of oil would rise or fall based on growth or shrinkage of the amount of oil in storage. Crude oil storage is not the only factor; storage considerations must include other petroleum liquids such as gas condensates and petroleum refined products.
Storage at Cushing currently is about 60 million barrels. Gulf Coast commercial storage is about 260 million barrels and the US Strategic Petroleum Reserve is about 680 million barrels. Total commercial US petroleum liquids storage is about 1.33 billion barrels; combined with the SPR, the total US storage is about 2 billion barrels. OECD storage, which includes the US, is about 3 billion barrels of SPR storage plus a little less than 3 billion of commercial storage for a total of almost 6 billion barrels. These figures are derived from many sources at different times and are subject to revision over a few months after they are first published. One cannot believe they are accurate on a given day to within 1.5% error.
Non-OECD data are even more problematic. Most of the countries in this group have no means to make accurate measurements and mostly do not want to. The largest country component in these data is China which has no interest in releasing accurate data on imports, exports, purchases, sales, consumption, or storage. Various estimates of the total storage of this group which, besides China, includes India, South Africa, and seaborne storage are around 2 billion barrels; that figure is estimated from various secondary sources.
If the Glut is defined as the amount of increased storage above historical levels it is very difficult to determine how much it is, where it is, or who owns it. As world daily consumption increases it is also expected that “normal” storage volumes will increase to maintain a certain number of days’ supply in storage.
Trading on the NYMEX is for contracts for 1000 barrels of West Texas Intermediate oil in a tank in Cushing, Oklahoma. Trading on the ICE is for 1000 barrels of Brent crude. The most accessible storage data for traders on the NYMEX or the ICE are for the oil storage terminal at Cushing, Oklahoma and US Gulf Coast storage followed by data for OECD countries. Press reports of “draws” or “builds” of Cushing storage can quickly influence trading prices on the NYMEX. The 60 million barrels of oil in storage at Cushing is roughly 0.75% of worldwide storage or 1.5% of worldwide commercial storage. Traders use Cushing data as a surrogate for variations of the worldwide Glut, evidently because it is the only data they have, not because it is a reliable indicator.
So to predict the price of oil tomorrow, next week, or the medium term or the long term, traders and analysts spend a lot of time micro-analyzing very bad data.
This takes our considerations back to the headlines quoted at the beginning of these comments and to the fact that the oil price started declining in August 2014 when the Federal Reserve stopped Quantitative Easing; not because of a supply “Glut”. It is also noted that when Quantitative Easing 2 started in November 2010 the oil price increased from a level fluctuating around $75 per bbl to about $100. When the Fed started QE the price went up and when it stopped QE the price went down.
I have attached a copy of a plot prepared by the Hein Energy Team, previously published by Enercom, showing the relationship between the dollar/ Euro ratio and the price of oil. The relationship between the value of the dollar and the price of oil is subject to endless debate as to causes and effects but the correlation is high, obviously better than coincidental, and raises many questions about The Glut.
Commentary.OilPricevsUS$-Euro.1
One cannot deny that short-term oil price volatility may be strongly affected by traders responding to variations of bad or irrelevant data. Typically, prices fluctuate daily and weekly over a $10 range around an average but the range may remain fairly constant over monthly and yearly terms. Significant changes to the high and low levels of those ranges seem to be responses to other influences, however. As one considers longer-term trends one cannot ignore the influence of financial market events, particularly actions of the Federal Reserve, which brings us back to the Petrodollar and its effects on the markets and government policy, if any, as discussed in the previous commentary.
I shall be traveling for the next 3 weeks or so but shall try to maintain these communications – but it may be a bit sporadic.