The Mystery of Petroleum Stocks

A number of days ago I started updating some datasets that had been left outdated for the lack of time. The jump in world petroleum extraction since early 2014 is obvious, but I wonder what happened to consumption in the meantime. The Energy Information Agency (EIA) in the US produces what I regard as the most reliable statistical dataset on petroleum available publicly. In times passed I used their worldwide stocks flow records to derive consumption from extraction figures. Unfortunately, the EIA stopped published this particular dataset.

This sent me on a quest for world petroleum stocks figures in order to estimate consumption. As it happens, there is at this stage no publicly available dataset on this matter, therefore any consumption estimates made presently have to rely on guesses. This post present the searching and guessing I did to arrive at my estimate.

On the rare occasions the mainstream or alternative media report stocks ,they exclusive refer to the figures issued by the International Energy Agency (IEA) contemplating only the stocks in the members of the OECD. In their latest issue of release of the Oil Market Report they IEA points to a figure of 4.6 Gb stored across the OECD at the end of 2015, having increased almost 300 Mb throughout the year. However, this figure only accounts for a fraction of the gap between extraction and consumption portrayed in the quarterly data issued by the IEA. In theory, the flows into world stocks should be much higher than what is reported for the OECD.

The first major problem in dealing with this mismatch is the inclusion of agricultural products (such as frying oil and schnapps) and other funny liquids in the IEA's statistics. These liquids are traded in different and markets and for entirely different prices. This is the reason why the quarterly "extraction" reported by the IEA is almost 20% higher than the petroleum figures reported by the EIA.

A couple of months ago John Kemp ran an article at Reuters precisely questioning this subject: Where has the oil gone? Missing barrels and market rebalancing. With the data available back then, Kemp concluded that since 2014 at least 550 Mb were unaccounted by the IEA, neither reported in stocks flows or in consumption. Kemp carefully let flow the idea that these figures were in error.

In their latest numbers, the IEA seems to have indeed closed this gap somewhat, but the total unaccounted figure remains quite large, as the table below shows.

Mb/d	Consumption	Extraction	Gap	Flow to OECD stocks	Unaccounted
2014Q1	91.96	92.14	0.18	0.24	-0.06
2014Q2	92.12	93.07	0.95	0.71	0.24
2014Q3	93.3	94.25	0.95	0.74	0.21
2014Q4	94.12	95.45	1.33	-0.11	1.44
2015Q1	93.6	95.07	1.47	0.89	0.58
2015Q2	94.12	96.26	2.14	1.03	1.11
2015Q3	95.67	96.98	1.31	0.77	0.54
2015Q4	95.48	97.23	1.75	0.42	1.33
				Comulative (Mb)	492

An obvious question is if these unaccounted barrels are being stored somewhere outside the OECD? The answer is yes, one has to rely on data fragments dropped in shady internet corners, but is seems that China is stocking up sizeable volumes of petroleum. In fact it could have well been responsible for half the gap between extraction and consumption reported by the IEA for the first quarter of 2016.

Bloomberg
Crude Tankers Bound for China Surge Amid Stockpiling Signals
Firat Kayakiran and Bill Lehane, 25-04-2016

China is hoarding crude at the fastest pace in at least a decade, filling inventories at a time when oil futures remain about 60 percent below where they were just two years ago. The nation added 787,000 barrels a day to stockpiles in the first quarter, the most for the period since at least 2004 when Bloomberg started calculations based on customs data. Its imports climbed in March from countries including Iran, Venezuela and Brazil.

With this information the guessing can now commence in haste. With two basic assumptions I estimate petroleum flows into China's stock: a) it accounted for as much as half of the gap reported by the IEA; and b) it never exceeded a rate of 0.75 Mb/d. Adding these estimates to the figures in the previous table results in the following:

Mb/d	Consumption	Extraction	Gap	Flow to OECD	Flow to China	Unaccounted
2014Q1	91.96	92.14	0.18	0.24	0.09	-0.15
2014Q2	92.12	93.07	0.95	0.71	0.47	-0.24
2014Q3	93.3	94.25	0.95	0.74	0.48	-0.26
2014Q4	94.12	95.45	1.33	-0.11	0.67	0.77
2015Q1	93.6	95.07	1.47	0.89	0.74	-0.16
2015Q2	94.12	96.26	2.14	1.03	0.75	0.36
2015Q3	95.67	96.98	1.31	0.77	0.66	-0.11
2015Q4	95.48	97.23	1.75	0.42	0.75	0.58
				Comulative (Mb)		72

Remarkably, the unaccounted barrels almost disappear. There are still 72 Mb left, but this could simply be fry oil stocks built by companies line Nando's.

Another interesting outcome of this estimate is the total petroleum stock amassed by China since 2014: 420 Mb/d, about the same figure reported by the IEA for the OECD. The following table presents comulative numbers since the beginning of 2014.

Mb	OECD	China	Total
2014Q1	22	8	30
2014Q2	87	52	139
2014Q3	154	95	249
2014Q4	144	156	300
2015Q1	225	223	448
2015Q2	319	291	610
2015Q3	389	351	740
2015Q4	427	419	846

With this petroleum stock flows estimate it is now possible to compute consumption. Mind here that this is a conservative estimate, since it counts on much of the unaccounted barrels to be ending up in China's stocks. The lower the stock flow estimate to China, the higher these consumption figures will be.

This estimate points to a consumption growth of at least 2.5% in the past 2 years. During the second half of 2015 consumption was persistently above 79 Mb/d. The most recent extraction figures published by the EIA point to a 0.6 Mb/d drop from December to January, down to 79.9 Mb/d. Although consumption usually falls somewhat at the beginning of the year, the gap between extraction and consumption could have well been already under 1 Mb/d in January.

Worldwide petroleum stock flows and consumption are thrown around almost daily by the media to interpret or drive what happens in the petroleum market. However, sound figures on these items are not publicly available, meaning that the media is at best guessing, at worst reasoning on wrong numbers. Opining is easy, objectively analysing the market not so much.

The Take Away

• No worldwide petroleum stocks dataset is publicly available today.


• The IEA has been overestimating the gap between extraction and consumption somewhere between 0.1 Mb/d and 0.7 Mb/d.


• Since 2014 China could have possibly stocked as much petroleum as the OECD put together.


• World petroleum consumption grew by at least 2 Mb/d since 2014.


• At the beginning of 2016, the gap between petroleum extraction and consumption was likely under 1 Mb/d.

Acknowledgements

I would like to thank Arthur Berman and Mason Inman for pointing appropriate directions in this issue.