August 04, 2009
Fatih Birol Of IEA: Peak Oil In 2020

Fatih Birol, chief economist of the OECD's International Energy Agency (IEA), has pulled in his prediction for when world oil production will peak to only 11 years from now.

In an interview with The Independent, Dr Birol said that the public and many governments appeared to be oblivious to the fact that the oil on which modern civilisation depends is running out far faster than previously predicted and that global production is likely to peak in about 10 years at least a decade earlier than most governments had estimated.

The reason for the change? The IEA has found that production from existing oil fields is dropping a lost faster than they predicted.

The IEA estimates that the decline in oil production in existing fields is now running at 6.7 per cent a year compared to the 3.7 per cent decline it had estimated in 2007, which it now acknowledges to be wrong.

The faster existing field production declines the faster new fields must be discovered and developed just to break even. The problem: world oil discovery peaked in 1965. In spite of all the technological advances since 1965 the discovery rate is less than a fifth the peak rate and less than a third of the yearly oil consumption rate.

The IEA's peak date is getting closer to that of some other notable observers. For example, oil industry analyst Charlie Maxwell predicts Peak Oil in 2015. Robert Rapier puts a 90% probability of a peak by 2013. Here is a consolidated collection of oil production forecasts.

The debate at this point revolves around how big will be the dislocations and drops in living standards? We can't know for sure. It depends on just how soon the peak comes, how fast the decline happens after it, and how quickly substitutes come along and at what costs. The timing of the peak is key. The later it comes the more technology we'll have available to deal with it. I hope for a later peak but fear an earlier one.

Update: A few months later it turns out the IEA was not allowed to report the more pessimistic earlier date for Peak Oil. The US government pressured the IEA to not report the conclusions they reached of a more rapid rate of oil field depletion.

Share |      Randall Parker, 2009 August 04 09:03 PM  Energy Peak Oil Adaptations

Engineer-Poet said at August 5, 2009 10:25 AM:

Peak by 2020 is an extremely safe prediction, given the observed facts:

  • Oil prices rose steadily at 30%/year from 2005 to 2008, making it ever more profitable to pump more.
  • Despite this increase in profitability, oil production stagnated; the price elasticity of production appeared to be roughly zero.
  • The recession caused by the 2008 price spike has slashed investment in oil production, so even the current level is extremely unlikely to be sustained.
The living standard we can enjoy henceforth depends how much we can squeeze out of a barrel.  If you can do everything without oil, oil production becomes as irrelevant to your standard of living as the production of spermaceti is to your household lighting.

Blick said at August 5, 2009 10:12 PM:

Oil is more likely to peak by 2050. But by then biofuels production will be 3 times current oil production, making the oil peak into something of a minor historical footnote. This much is obvious, but should be pointed out to those not paying attention.

Randall Parker said at August 6, 2009 1:35 AM:


If what you say is obvious then why are a lot of people paying attention and yet believing that Peak Oil comes either very soon or has already happened? Why does the CEO of the French oil giant Total saying that oil will never reach 100 million barrels per day? Why is Fatih Birol (who is paying far more attention than you) saying it is happening in 2020? Why is Charlie Maxwell (who has been paying attention for over 50 years) saying it happens by 2015?

Jens said at August 6, 2009 3:27 AM:

I thought it peaked last year? In march 2008 production was around 83 mill barrels per day and consumption around 85 mill barrels per day. It's gone down since, partly because of recession of course, but still.

Engineer-Poet said at August 6, 2009 7:44 AM:

In case nobody got my point, I'm calling the all-time peak last year in 2008.  More to the point, Jens, the peak last year was "all liquids".  However, all liquids are not created equal.  Crude oil is declining, while the fraction which is increasing is natural-gas plant liquids, largely C2-C4 hydrocarbons.  These have far less energy per barrel than crude, so the actual energy production from these liquids is falling.

Blick:  Please tell us where you would get 250 million bbl/day of biofuels.  Even if this is 100% ethanol (one of the least carbon-intensive fuels), 250 mmbbl/day is 31 millon tonnes/day of ethanol containing ~16 million tonnes/day of carbon.  At a conversion ratio of 100 gallons per short ton of biomass, this would take 95 million metric tons of biomass per day, 35 billon metric tons per year.

The feasible production of new biomass in the USA is approximately ONE billion tons per year.  What you describe requires breakthroughs on a scale equivalent to the blue-sky "too cheap to meter" speculation about nuclear power.

Nick G said at August 7, 2009 9:13 AM:


My view on peak isn't terribly different than yours, I suspect, but it's worth noting that high prices did indeed manage to push oil production off it's several-year plateau to a peak in 2008 - see Rembrandt's last Oilwatch Monthly July 2009

Markets still matter.

Randall Parker said at August 8, 2009 7:44 PM:

Nick G,

The number of price doublings that it took to make a new peak in 2008 that beat the 2005 peak tells us that the cost of oil extraction is getting much higher. This is consistent with industry statements. I just read a statement by a Schlumberger guy that the market needs $80 per barrel oil to go back up in drilling activity.

The international oil majors need much higher prices in order to maintain production.

But, Neil McMahon of Sanford Bernstein calculates, the majors would struggle to make operating cash flow cover the cost of replacing reserves and increasing production while also paying dividends, even if oil averaged $80 per barrel and natural gas averaged $9 per thousand cubic feet in 2010. Both numbers are significantly above the Street's consensus forecasts.

Still, at least in North America oil drilling activity hasn't dropped much

The most recent drilling rig counts by Baker Hughes Inc. show there were 943 operating in the U. S. on July 24, down 1,014 from the same time in 2008. In Canada, there were 180, off 255 from 12 months earlier.

In Mexico, there were 124 rigs operating in June, according to Baker Hughes, compared with 112 in the same month of 2008 and 90 in June 2007.

In its report, CIBC estimates that activity in Chicontepec region could keep 70 rigs busy on a constant basis, adding it expects the total on the ground to increase to 65 by the end of 2009.

Nick G said at August 9, 2009 5:08 PM:

hmmm. Contradictory info. here are some thoughts:

The number of price doublings that it took to make a new peak in 2008 that beat the 2005 peak tells us that the cost of oil extraction is getting much higher.

Don't forget the lag time for medium & large oil projects. The peak prices of 2008 were around for far too short a time to mean anything. Really, the peak in 2008 was a response to prices in 2006, and some of 2007.

Also, the largest oil companies, especially Exxon-Mobil, maintain pretty steady "green-light" numbers. If the price spikes or crashes, it doesn't have that large an impact: they have a long-term approach.

Many of these numbers seem to conflate oil & gas, which have diverged. At this point, in theory no one should be drilling for gas in NA at all, given that storage is going to max out before the end of the injection season.

We really need numbers that separate oil & gas rig statistics. The story about rig counts suggests that they've dropped more than 50% in NA, but most of the drop will be gas drilling.

Finally, what % of oil is produced by the IOM's? They only have about 5% of the reserves, IIRC.

Nick G said at August 9, 2009 6:11 PM:

A few more thoughts.

The divergence between WTI and most oil in export markets means that the headline oil price is overstated.

More importantly, the fall in the dollar means that 2/3 of oil buyers in the world pay 2/3 as much as Americans.

So, prices in 2008, as well as now, weren't nearly as high for most oil producers as they seemed to us in the US.

Demesure said at August 10, 2009 1:49 PM:

It's reassuring that peak-oil will (maybe) only occur in 2020 (nice round number, easy to remember for dumb people) given that peak-oil has been announced for 1990, 1991, 1992, ... 2000, 2001, 2003 ... 2008, 2009 by peak-oilers.
What a relief !

Nick G said at August 10, 2009 2:07 PM:


who announced Peak Oil for 1990? Do you have a link?

Tom S said at August 12, 2009 1:05 PM:

It really doesn't matter what the numerical peak is or when it occurs. More important is when demand (without supply constraints) exceeds supply capacity. It is relatively easy to forecast future demand given that, since 1983, world per capita oil consumption has been 4.53 B/P/Y with a standard deviation of 0.094 B/P/Y. Using this we can forecast 2015 oil consumption (demand) at between 88.4 MB/D and 92.2 MB/D. Given Fatih Birol's recent remarks,, that could be problematic.

Nick G said at August 12, 2009 2:22 PM:


Some people call that "Peak-Lite".

It seems to me that we already hit that point around, say, 2006, when prices started to rise.

To me, the interesting question is when we hit Peak Demand: the point at which 1) substitutes like PHEV/EVs start to hit economies of scale and become decisively cheaper than oil, and 2) substitutes start to really ramp up enough to reduce oil consumption.

Post a comment
Name (not anon or anonymous):
Email Address:
Remember info?

Go Read More Posts On FuturePundit
Site Traffic Info
The contents of this site are copyright