Thursday, January 29, 2015

Interesting things in world oil balance tables

“Great part of the information obtained in War is contradictory, a still greater part is false, and by far the greatest part is of a doubtful character” - Carl von Clausewitz.

If Clausewitz were alive, he would perhaps say the same thing for world oil statistics. In my previous post I tried to demonstrate that you can come up with a completely different picture if you take world oil supply estimates from one source and world oil demand figures from another. In this post I would like to draw your attention to some interesting things in world oil balance tables.

One of the strangest items in the IEA’s OMR is “Miscellaneous to balance” that appears at the IEA’s famous Table 1- WORLD OIL SUPPLY AND DEMAND. The footnote on the table defines it as “Includes changes in non-reported stocks in OECD and non-OECD areas.” It was 1.0 mbd in 2Q14 (oh well, also -1.4 mbd in 3Q13). 1 million barrels per day of unknown oil (to me) in 2Q14 but nobody talks about it.

Do you remember the “missing barrels” phenomenon (try a Google search you will understand what I mean)? The late Matthew R. Simmons had written a few good articles (including this one). (See also my posts on May 24, 2006 What Oil Market Fundamentals? And Missing Barrels are Back on 13 Sept 2005.)

At that time I was working at the IEA and I was asked to investigate the reasons. I came up with some reasons and suggestions. They were simple. This is why they were mostly rejected. One of those was on conversion factors. It was ignored at that time but was partially applied years later.

Today, I say it again: conversion factors (from metric ton to barrels) are the Achilles' heel of the world oil supply-demand tables. Note that the figures in the IEA’s oil supply-demand tables are expressed in thousand (or million) barrels/day which are converted from metric ton using volume to mass ratios in barrels/ton. OECD Member countries report their oil statistics in metric tons. All those figures are then converted to barrels at the IEA by using…, well that’s the problem.

Crude oil barrel-ton conversion factors usually vary by country but interestingly not by year (except for some rare exceptions). See oil documentation. For instance, crude oil production in Norway and UK go down but barrel-ton conversion factor has always been constant (7.37). Anyone in oil business knows that oil gets heavier especially during the post peak period.

Barrel-ton conversion factor may not be too important. What I have been saying for more than a decade now is that it is chemical and statistical crime to keep barrel-ton conversion factors constant for all years and for all countries (there are very few exceptions).

Coincidentally I found out that in the 2014 Edition of the IEA’s Energy Statistics barrel-ton conversion factor for Motor Gasoline is different from the previous edition. Motor Gasoline here refers to gasoline excluding biofuels. Previously it was 8.25 for almost all OECD countries and 7.93 for nearly all non-OECD countries. This means 1 metric ton of gasoline equals to 8.25 barrels in OECD and 7.93 barrels in non-OECD countries. In the 2014 Edition this factor was changed to 8.53, for most of the countries in the world, OECD and no-OECD alike. In other words the conversion factor was increased by %3 to 7%. (I skip here why and whether it makes sense).

Since the gasoline consumption by country in metric tons hasn’t changed in the IEA’s energy statistics database, consumption figures in barrels should have changed by the same magnitude of the change in conversion factor. According to my rough calculations world gasoline consumption in terms of barrels should have been increased by around 5%. This was not the case, however. World gasoline consumption in barrels remained the same. The reason? I don’t know. I use metric ton figures for my work anyway. So, no impact on my work. But it is highly likely that the IEA’s energy statistics (hence OMR) understates the world gasoline demand in barrels, and hence the world total liquids demand by around 5%. Yes, small in percentage but it makes around 1.5 mbd. Maybe I am doing something wrong. Please correct me.

Now let me jump to another issue: Processing Gains. Compared to the year 2000, in 2013 an additional amount of some14 mbd of petroleum products were produced in the refineries globally. But over the same period the refinery (or processing) gains increased by only 0.4 mbd (from 1.8 mbd in 2000 to 2.2 mbd in 2013). Refinery gains worldwide in 2013 are shown as 2.1 or 2.2 mbd in OPEC and IEA oil balance tables. We know that in 2013 the processing gain in the US was 6.9% or 1.07 mbd for total net refinery production of 12.1 mbd. This means that the US accounts for half of the global processing gains despite the fact that only 15% of world refined products are produced in the US. Does this make sense? Not to me.

Conversion factors affect the US EIA and OPEC as well. Remember, not all countries in the world report their oil supply and demand figures in barrels. The original unit must be converted into barrels. Processing gains is also another item both EIA and OPEC should look into very carefully, and stop copying the IEA’s figures.

If you are interested in conversion factors see for instance an article I wrote 11 years ago (Conversion factors and oil statistics, Energy Policy, Jan 2004).


In sum, oil industry needs good data. This requires putting more money and resources on statistics. Scrapping the budgets in national statistics offices may lead to the erosion of data quality and coverage, particularly on oil demand.