(Note: Commentaries do not necessarily represent ASPO-USA’s positions; they are personal statements and observations by informed commentators.)

It is difficult to make predictions and one tends to shy away from criticizing other people’s incorrect forecasts. Next time it might be you making an error. However some people and organizations have such a high level of credibility and profile that they are trusted by politicians and captains of industry to deliver solid information.

We know these institutions influence decisions that will impact global development. If these institutions year after year deliver bad data, and if they at the same time refuse to accept that something is wrong with their forecasting models, it is actually an important service to point this out. Important decision makers in politics and industry should know that what these institutions say about the future can not be relied upon as a solid basis for policymaking.

Track record of the IEA in making forecasts.

Each year in the July issue of the Monthly Oil Report, the International Energy Agency in Paris (IEA) publishes their first forecast for next year’s supply and demand. The geographic area where the IEA should have the best chance of making accurate predictions must be non-Opec ex FSU. In this area there is limited civil upheaval and political interference with the oil industry.

Since their forecast made in 2002, it turns out they have increasingly overestimated production growth each consecutive year.

Table 1: Production Non-Opec excluding FSU

In 1000 b/d 2003

IEA forecast 38,900 39,600 39,600 40,200 38,650(1)

Actual 38,800 38,900 38,700 38,800 37,940(2)

Mistake 100

1: For 2007 the figure excludes Angola which is now an OPEC member
2: This figure is as pr. the February report of the IEA. Further downward revisions are likely and the 700,000 b/d will therefore most likely increase. In February 2006 the 2006 estimate had been revised down with 1 million b/d and the year ended at 1.4 million b/d. Adjusting for Angola IEA today says there will be zero growth from 2003 to 2007.

Just after the summer of 2006 a representative of the IEA was in Oslo presenting their 5-year outlook. In his presentation the speaker mentioned that in 2005 their forecast had been off, but that was due to one-time events like hurricanes and political upheavals. The speaker was challenged on the statement that 2005 was an unusual year. The trend from 2003 was pointed out, and at the time of making the presentation, the first half of 2006 was already 1.2 million b/d below IEA’s original forecast. Could there be something wrong with the model and assumptions used by the IEA? The answer was without hesitation: There was nothing wrong with the model. In the second half of 2006 we would see a surge in production which would prove their model right.

As the fall progressed the speaker from the IEA must have had a sinking feeling in his stomach watching month by month 2006 being revised downwards. Not only did the surge not happen, but the second half of the year was actually performing even worse than the first half bringing the total in downward revisions to 1.4 million b/d.

As a blessing in disguise their demand forecast was almost as much off. 2006 global demand growth turned out to be 1.2 million b/d lower than forecast. Had the original forecast been accurate we can only guess at what the oil price would have been last summer.

As for the 2007 forecast the IEA started out in July 2006 with a growth in the non-Opec ex FSU area (excluding Angola, which has joined Opec) of 860,000 bpd. By February 2007 the growth in supply from 2006 to 2007 had been reduced to 580,000 bpd. However as actual production in 2006 was lower than predicted in July, the February forecast for total production was reduced with 700,000 b/d from the July forecast.

Upon being questioned the IEA explains that their supply model is based on a field-by-field, region-by-region basis. They gather information from the operators of the oilfields regarding a field’s (or region’s) expected performance. They admit to tweaking the numbers based on their own judgment. But as they said in an e-mail: “we would be misleading our subscribers if we took an arbitrary slice out of non-OPEC growth because of a subjective judgment”.

My question is: Why not? It is evident that the information they receive from the operators is too optimistic. Either the information they receive must be tweaked some more, or if they do not want to point the finger at anyone, one could make a statement such as: Based on years of experience we are convinced that the information we receive in the aggregate is overoptimistic. We therefore reduce the global supply number with x% to reach a realistic forecast. Unless such an adjustment is made, either in the aggregate or on a field-by-field basis, the forecast that the IEA is sending out is actually what is misleading.

Without going into details it looks like the two big culprits for the current problems are:

1. Global decline rates are higher than assumed. If we on a global basis are wrong on the managed decline rate with 1%, it represents lost production of 850,000 b/d annually.

2. Projects tend to suffer more slippage than expected. In a report from Simmons & Company, a 10% added project slippage will in 2010 deduct 1.4 mil b/d from global production capacity.


Cambridge Energy Research Associates (CERA) published in 2005 a very hotly debated study claiming that the global production capacity of liquids will grow 16.4 million b/d through 2010. If the forecast is right, oil prices will fall sharply. If so, the prudent decision by oil companies will be to invest less in E&P [exploration and production]. If the captains of industry and government listen to CERA and act accordingly, and it turns out they are wrong, the global economic well being will be severely hurt as the only way to balance supply with demand will be increasing prices.

If one is of a conspiratorial mind, one could point out that CERA is a profit-making consultancy firm whose most important customers are the big oil companies like Exxon and Aramco. One could also think that if CERA is right in their forecasts, then the oil companies do well in holding back investments. Should CERA be wrong, the oil companies will still do well, as the oil price will certainly be high. So, if you were an oil company executive, what would you like to have your consultant tell you?

In August of 2006 CERA was bold enough to make an update of their own study. The update was presented to the readers and the public as if the 2005 study was right and the expansion is on its way. Let us look at the figures:

Table 2: CERA Global liquids capacity development


May 2005 report 85,080 87,850 90,650 93,760 96,490 101,520

August 2006 84,730 86,260 88,740 91,620 94,830 102,050

Change in 15 months -350

IEA as pr February 07 83,200 84,500 85,300 86,000

Difference CERA – IEA 1,530

Comparing IEA and CERA is to a certain extent comparing apples and oranges, as CERA is dealing in liquids capacity while the IEA is dealing in actual production. One should expect the difference within this timeframe to stay fairly constant unless one is of the opinion that there has been significant growth in spare capacity. Particularly in the summer of 2006 one would tend to say that there was no real spare capacity. How can then CERA explain the divergence in the two lines?

In a handout from a CERA presentation last fall we also find that CERA’s estimate for growth in non-Opec capacity was 1 million b/d for 2006 and 1,7 million b/d for 2007. According to the IEA the growth in supply in 2006 was 600,000 b/d and so far their estimate for 2007 is 1.3 million b/d (including Angola), an estimate we expect to be further reduced. CERA is at least 800,000 b/d behind their own schedule. For CERA a 2007 revision of their 2006 revision will definitely be necessary.

Looking at the table we can make some interesting observations. The forecast of May 2005 has turned out to be terribly wrong. Let us start with 2004, a year already 5 months old at the time of the report. The global production in that year was adjusted upwards with 200,000 b/d by the IEA from the May 2005 issue to the August 2006 issue. In the same period CERA reduced their capacity figure for 2004 with 350,000 b/d.

As for 2005, 2006 and 2007 we see that their capacity expansion is greatly reduced in the 2006 update. One could say they have sobered up to reality. However, in order to be able to say that they “stay the course” they play catch up in 2009 and 2010 to the extent that for 2010 the expected capacity in the 2006 update is 500,000 b/d ABOVE their 2005 report.

It means that capacity grows with 2.3% p.a. in the two year period from 2004 to 2006 but with 4 % p.a. in the two year period from 2008 to 2010. We already know that the IEA figures for 2006 and 2007 was downgraded considerably from August last year. Consequently, should CERA make similar revisions and still reach their 2010 target, the growth rate must be well above 4% p.a. in 2008 to 2010. When was the last time that happened? The 1970’s?

This notion also flies in the face of the generally accepted fact that non-Opec production growth will fall towards the end of this decade.

When CERA in 2005 missed so terribly with their forecast for the next 3 years, how can we believe any of what they say about 2009 and 2010? Their answer is, trust us, we have this phenomenal field-by-field databank which allows us an accuracy in our predictions that nobody else can compete with.

Let us put this phenomenal accuracy in their forecasts to the test. Let us look at an oil province close to me, the Norwegian continental shelf.

In the 2005 study CERA said that Norway would produce 3,590,000 b/d of liquids in 2006, up from 3,360,000 in 2004. Here is what their table looks like:

Table 3: CERA forecast for Norwegian Continental Shelf

In 1000’s b/d 2004

May 2005 report 3,360 3,340 3,590 3,460 3,330 3,030

August 2006 3,300 3,170 3,160 3,170 3,270 3,150

Change 15 months -60

IEA as pr. Feb.2007 3,190 2,970 2,780 2,700

Difference CERA-IEA 110

NPD 2,780 2,600 2,600 2,400

Difference CERA-NPD 570

The NPD (Norwegian Petroleum Directorate) figures are from January 2007. The IEA has not yet fully taken this into account, which may indicate future IEA revisions. As to comparing with CERA it is more realistic for Norway than on a Global basis, as CERA will hardly claim any spare capacity in Norway. As we can see, even with CERA’s August downward revisions, they are far from reality. Again they have this mystical upturn in 2010 compared with their 2005 report.

From the table we can see that in two years, from 2004 to 2006, instead of having a growth in production of 230,000 bpd, Norway has seen a drop in production of 410,000 bpd. A total error
of 640,000 b/d in less than two years ! For 2007 NPD expects 2.6 million b/d in liquids production and in 2010: 2.4 million b/d. By then the gap to CERA has grown to 750,000 b/d.

Those that pay CERA good money to get access to their data should ask themselves: CERA claims to have all this detailed field-by-field data which makes their forecasts so trustworthy. What is the value of this field-by-field data when they can be this much wrong in such a short time, in the country in the world with the best field-by-field data available in the public domain? When they are this wrong in Norway, is there any reason to believe they are less off the mark in Angola or Kazakhstan? Should I really spend good money on their reports, and why should I rely on them?

Spare Capacity

The IEA has made presentations of their 5-year outlook and journalists are looking for sound bites. One such sound bite is that we will have 6.8 million b/d in spare capacity in 2010. The press reports that this will bring oil prices down at the end of the decade. Interestingly enough, when CERA makes similar presentations they have 7 to 8 million b/d of spare capacity in 2010.

What is really spare capacity?

According to the IEA the definition is production capacity which can be brought on stream within 30 days and be kept producing for at least 90 days.

If you look at the August 2006 report from the IEA, their spare capacity figure was 3 million b/d. They explain in the table that if one deducts the notional spare capacity in Iraq, Nigeria, Venezuela and Indonesia the spare capacity is 2 million b/d. It would certainly take more than a political decision to make these countries increase their production.

Most of the remaining 2 million b/d is oil which is so heavy that it can not be refined with today’s refining capacity. If it can not be used, is it really spare capacity? There may be peace in Iraq, one day, and we may expand refinery capacity sufficiently to refine the 2 million b/d of heavy oil, but not in 30 days — maybe not for years. Is it not then deceiving to call it spare capacity? The press will 9 out of 10 times fall for the easy interpretation and report that we have indeed spare capacity should the world need it.

The 5-year outlook from the IEA and future spare capacity.

If you look at the 5-year IEA forecast published in July last year on page 7 you can see that in 2010 we are set to have a spare capacity of 6.8 million b/d. That sounds comforting and should guarantee lower oil prices. However looking at 2006 in the same table we see that the spare capacity for 2006 was 4.2 million b/d. This is more than reported in the monthly reports. Most of the extra is stock change. It is not clear if it includes all stock change or only what exceeds what is necessary to keep stocks at level days of forward coverage. Regardless it looks too high and one may wonder if it also includes some of IEA’s infamous “missing barrels”.

What they say in reality is that spare capacity over the next 4 years will grow with 2.6 million b/d, provided everything goes as planned. And we already know that the rest of 2006 definitely did not go as IEA expected in July.

Interestingly enough the figure for non-Opec production 2010 was reduced from 56 million b/d in the July 5-year outlook to 53.4 million b/d in the 2006 World Energy outlook. Demand was reduced only with 400,000 b/d in the same period. It seems that the 2.6 million b/d build in spare capacity from here to 2010 stated in their July report is already reduced to 400,000 b/d. And this is assuming all goes well as per IEA’s expectations. With IEA’s track record for overestimating supply growth it is surprising that this has not created a reasonable level of panic among energy decision makers.

We have one unknown factor in this equation. Will the refinery capacity for heavy oil increase so that it can handle the “spare capacity” of heavy oil between now and 2010? If the answer is “yes”, we may see an additional 1.5 to 2 million b/d available on the market.

CERA says that global spare capacity in 2010 will be between 6 and 8 million b/d. In their hand out they invite us to send comments and questions to the lecturer. I have via e-mail several times tried to ask how much spare capacity CERA counted in 2006 when making that forecast. For some reason they can not find time to answer. However, in their hand-out they stated this fall that about 1.8 million b/d is held off the market for various reasons (Iraq, Nigeria, Venezuela and the US), and if they treat the heavy oil like the IEA does, they probably operate with at least the same spare capacity as the IEA, possibly almost 1 million b/d more.

The comfort of the 6 to 8 million b/d in spare capacity we are told by CERA we will have in 2010, of course depends on whether they say:

  • In reality we had very limited spare capacity in 2006 but it will grow to 6 to 8 million b/d in 2010, or they say
  • In 2006 we had 4-5 million b/d in spare capacity and the spare capacity will grow with 1 to 3 million b/d through 2010.

Somehow the latter statement sounds far less reassuring than: “Relax. We have 16.4 million b/d of new capacity coming on stream by the end of the decade.” For some reason it is the latter statement the press is usually quoting.

CERA has made it their task in life to debunk the peak oil myth. They claim that the resource base is plentiful and technology will solve all problems. Somehow many people listen. Would they listen as much if they knew the huge errors CERA has been making? How their forecasts, when analyzed and compared with other sources, look completely unrealistic? And that even based on their own forecast the spare capacity will actually only grow with 1 to 3 million b/d? If the global oil industry cannot do better than that after a decade of unprecedented oil prices, far above anybody’s marginal cost, it means something is terribly wrong with our ability to grow production. In this situation one does not need King Hubbard to understand that “Houston, we have a problem.”

Aage Figenschou is a lawyer who lives in Oslo and is a consultant to a variety of public and private companies in the oil & gas and shipping industries.