In my last post, I noted how even the IMF has come to address the issue of Peak Oil, most explicitly in the April 2011 edition of their flagship publication World Economic World (here). From a risk perspective, the first take-away from the report is that the very inelastic short-term supply and demand curves for oil make the world economy (and individual wealth) highly vulnerable to any geopolitical disturbance in oil supply. The second take-away is that long-term demand is also highly inelastic, suggesting any substitution away from oil is problematic (for example, it would take decades to replace our petrol and diesel based car fleet with an electric-based alternative). However, neither of these conclusions directly address the issue of whether oil production could peak.
On the longer-term supply side, the central scenario in the IMF report is for an annual average growth rate in oil production of 1.5%, while the the alternative ‘Peak Oil’ scenario (called Scenario 2 in the report) is for a decline of 2% per annum. How exactly these figures have been arrived at is left somewhat vague. The supply growth scenario appears an extrapolation of recent trends, while also being broadly consistent with the forecasts of the International Energy Agency (IEA). For the contraction scenario, a reference is given to a paper by Sorrell et al but no further details are provided. In other words, the IMF deftly avoids going into the peak oil controversy but just plucks out a couple of scenarios: 1) consensus oil cornucopia and 2) a mild oil descent.
The lack of an explicit treatment of the long-term supply side is somewhat puzzling, given that the IMF’s study feels able to put numbers onto the demand side (albeit with large caveats). In short, no light is shed on how price can introduce more capital expenditure and thus call forth more oil supply over the longer term and no information is given on how the supply curve moves through time due to the impact of technology.
Nonetheless, if you access the paper led by Steve Sorrell that is referenced by the IMF (here) you can find a treasure trove of information about a range of possible long-term oil supply outcomes, both optimistic and pessimistic. Unfortunately, the article is behind a pay wall (and costs $19.95 through the Science Direct website) but most of the contents are freely available in other publications Steve Sorrell has co-authored—particularly the The Global Oil Depletion Report from the UK Energy Centre—and in video and powerpoint presentations.
If you want to understand the main points Sorrell wishes to communicate directly, then I suggest you watch the video below (as the powerpoint charts are barely visible in the video, it is worth running this Powerpoint presentation concurrently as most of the slides overlap). For those time constrained, I have pulled out the most interesting points and charts in the rest of the post.
If I had to sum up Sorrell’s key messages, they would be these:
- Consensus reserve estimates of oil reserves may be accurate (thus falsifying the claims of high profile Peak Oil exponents such as Colin Campbell), however this makes only a relatively small difference to the timing of a peak in production.
- The critical issue is how easily (and cheaply) we can access resources rather than their ultimate size.
- There is a significant risk that oil production will peak before 2020.
- The assumption of a peak beyond 2030 appears at best optimistic and at worst implausible.
Sorrell starts his analysis by defining conventional oil as a combination of crude oil, condensates and natural gas liquids.
Sorrell notes that true unconventional oil currently accounts for only 3% of total ‘all liquids’ production and is projected by the IEA to only be a little over 10% by 2030 (the IEA’s Current Policies scenario sees 11.3 million barrels per day of unconventional oil production in 2030 out of a total of 103.9 mbd in their 2011 World Energy Outlook report). The IEA defines unconventional liquids to include biofuels plus unconventional oil which includes oil sands, extra heavy oil, gas to liquids (which is different from natural gas liquids), coal to liquids and kerogen oil.
Against this background, Sorrell stresses that if conventional oil depletes more quickly than expected, non-conventional oil will not be able to make up the difference over a 20 year time span. He then focuses on the supply dynamics of conventional oil, particularly on the physical constraints on production, and notes the following points:
- Each individual oil fields follows a pattern of rise, peak, plateau and fall
- Most production in a region originates from larger fields
- Larger fields are relatively easy to find than smaller ones and are thus generally found first
The UK oil industry is used by Sorrell as a typical example. The largest fields such as Brent and Forties were found in the early stages of exploration. Further, the much smaller fields most recently discovered both peak earlier and decline quicker as well as produce less oil overall. Note that the recent numbers from the latest BP Statistical View of World Energy show UK oil production at 1.3 mb/d in 2010, down from 2.7 mb/d in 2000.
Economists please note (and I am one by training): during the period of UK oil production decline, the oil price has risen radically and deep sea drilling and extraction techniques have continued to improve. Nonetheless, the decline in production has been impervious to both price and technological trends. What is more, numerous other countries are now in the position of recording relentless output declines.
Sorrell then looks at the concept of ultimately recoverable resource (URR). In his words
The URR is the sum of cumulative discoveries, future reserve growth at known fields and the volume of oil estimated to be economically recoverable from undiscovered fields—commonly termed the yet-to-find (YTF).
Current estimates for URR are between 2,000 and 4,300 giga barrels (Gb), with giga meaning billion. If we take out what has already been produced, that leaves us with between 870 Gb and 3,170 Gb left to recover. (The IEA in its 2011 WEO report puts recoverable conventional oil resource at 2,800 Gb, and thus a URR of around 4,000 Gb.) Sorrel then goes on to make the critical observation that the year of peaking is remarkably insensitive to the actual URR:
Now you can alter the shape of the curve to delay the peak, principally by a) decelerating your oil production ramp up, or b) putting your peak off but at the expense of having oil production fall off a cliff at some point in the future. But neither action is a free lunch economically.
Sorrell then went on to analyse 14 forecasts of oil production out to 2030. Of these 14, five forecast no peak before 2030; these forecasts came from the IEA, OPEC, US Energy Information Administration (EIA), ExxonMobil and Meling (Statoil-hydro). The non-peak forecasts arrived at their conclusions through a combination of having higher URR estimates and assuming larger post peak declines:
The next critical question then is how realistic are the no-peak-before-2030 forecasts? Sorrell implies that such forecasts should not be judged as central scenarios but rather as at the edge of all possible outcomes for a number of reasons.
First, historical experience suggests that production curves are asymmetric—but not to the right of the peak but rather to the left. In other words, we see a fast ramp-up and then slow decline. Sorrell notes that of the 37 countries that have already experienced a peak, the peak took place at 26% (production weighted) of their URR (ultimate recoverable reserves). The more optimistic forecasts—like those of the IEA, EIA and OPEC—are implicitly looking for a peak at a much higher percentage of URR.
Second, the more optimistic studies assume that discovery trends will remain in place despite the fact that they have been on a long-term downward decline (although the IMF’s New Policies Scenario has quite conservative discovery rates of 8 Gb per year through 2035).
Third, many forecasts such as those by the IEA suggest that oil can be extracted from the newly discovered fields at a rate that has no historical precedent.
Following on from this analysis, Sorrell and his co-authors ended the paper cited by the IMF with this statement:
Given these complexities, we suggest that there is a significant risk of a peak in conventional oil production before 2020. At present, most OECD governments are failing to give serious consideration to this risk, despite its potentially far-reaching consequences.
From a risk perspective, this appears very sensible. We just don’t know when the actual conventional oil peak will take place and attaching a single point estimate appears a futile exercise. However, we do know that it could appear early or late. If early, this would imply that we have very little time to put in pace either non-conventional supplies or renewables. The result of a supply side conventional oil shock could thus inflict a major blow to global GDP (and, potentially, global political institutions).