Welcome to the age of diminishing returns

Thursday, February 9, 2012

The fate of new truths: peak oil appears on "Nature"

"It is the customary fate of new truths to begin as heresies and to end as superstitions" (Thomas Henry Huxley, 1880) Above: a figure from the article by James Murray and David King published on Nature, 26 Jan 2012, vol 481, p. 435

With the publication of a prominent article on "Nature" in January 2012, the concept of "Peak Oil" has made another step forward in the debate on resource depletion. This article has made me rethink of the past ten years of work that I did as a member of ASPO, the association for the study of peak oil. Were we right with our prediction of impending peak oil? In a sense, yes, but the crystal ball is always foggy and it cannot be otherwise. The ASPO predictions were basically right but, as all predictions, they were approximate.

Working with a simplified model based on Hubbert's early work of the 1950s, the founders of ASPO, Colin Campbell and Jean Laherrere, proposed in 1998 that the future of oil production would have followed a curve that was to peak at some moment between 2005 and 2010, to decline afterwards. Embedded within the Hubbert model was the concept that the gradually increasing costs of extraction would reduce the profits of the industry and force it to reduce investments.  

As a "first order" model, the Hubbert one is not bad and the ASPO models caught very well the problems that the oil industry was going to face. From 2004 onward, prices have shot up a levels that have changed forever the oil market. But oil production, intended as "all liquids" (that is, including oil from tar sands, biofuels, etc) didn't show a well defined peak, nor the decline that the Hubbert model predicted. Stubbornly, production has refused to decline and it may even be showing a modest increase in recent times. That doesn't make the model wrong: as all models, it is an approximation of reality. 

The "peak oil" concept has been often criticized on the basis of a classic idea in economic science: that prices mediate between demand and offer. Hence, oil prices should define what should be counted as "reserves", intended as something that can, and will be, extracted. High prices should lead to more reserves and we would never run out of anything. It turns out that this criticism was not wrong, although not right, either, and its consequences were perhaps unexpected even for those who proposed it. When scarcity started to be felt in the oil market, the price correction mechanism kicked in. Prices rose and, according to the standard economic theory, that should have stimulated production. It did, in part, but with crude oil the mechanism has become a rat race. The more high prices made production profitable, the more production costs rose. That's where we hit the ceiling. 

This mechanism is very nicely caught by Murray and King in their article in Nature. The figure reported at the beginning of this post shows it very clearly. Over a certain price, production doesn't respond any more. It becomes "inelastic". The graph has to be read taking into account the temporal evolution of both prices and production: very high prices are a recent phenomenon and what we see is what I called the rat race. Even with increasing oil prices, the best that the industry can do is to keep constant the production of combustible liquids. 

So, we are seeing that the price mechanism may slow down the expected production decline, but at the price of causing all sorts of problems. With high prices, the world's economy must allocate more and more resources to oil extraction and these resources must come from somewhere. Since the economy doesn't grow any more, keeping oil production constant means that some sectors must shrink and that is not without pain. Much of the present political turmoil in poor countries, for instance, is due to the high prices of food, in turn related to the high cost of oil. And, with prices so high, we see the perverse effect that producers can afford to consume more but, as a result, less oil is left for importers. In a sense, many importing countries have already passed their peak oil.

As Thomas Huxley said long ago, it is the customary fate of new truths to begin as heresies and to end as superstitions. Peak oil surely began as a superstition and it is still considered as such in some circles. But, with the events of the past few years, it is also attaining the status of truth, as shown by the article by Murray and King, who have clearly understood what lies at the basis of the idea. In some sense, however, peak oil is also taking some elements of a superstition, since it fails to take into account the price mechanism. In the end, reality might be better described by something like the "Seneca model" which takes into account second order effects and that predicts a production plateau followed by a sharp decline. Also this model may be a heresy, right now, but one day may become truth and later on a superstition. As always, the future is never what it used to be. 


The paper by Murray and King on Nature is here (behind a paywall)
A summary can be found on Scientific American here
A comment on the New York Times is here
A criticism by Michael Levi can be found here
And a defense by Mason Inman, here


Note 1. I completely agree with the approach of Murray and King regarding the relation of peak oil and climate change. It is true that the two problems are strictly related and that they should be tackled together. However, I also think that the authors should have been more careful in the way they presented this issue. At the start of the article, they say,  "....continuing debates about the quality of climate-change science and doubts about the scale of negative environmental impacts have held back political action against rising greenhouse gas emissions. But there is a potentially more persuasive argument for lowering global emissions: the impact of dwindling oil supplies on the economy." Consider the number of conspiracy theorists around, this paragraph will surely be seen it as "proof" that peak oil is a hoax created by the evil oil companies in order to force customers to pay higher prices for gasoline. Besides, it makes no sense in my opinion to say that scarcity is a good argument to convince people to consume less. It would be, if people behaved rationally, but most people don't. It reminds me of an experience I had some time ago, when I presented the peak oil case to a rich financial tycoon. He answered to me with something like, "I think you are right. So, I guess I should buy myself a new Ferrari and consume as much as I can, while I can."

Note 2. Italian readers of this blog may be interested in this paragraph from Murray and King's paper. It think it is right on target.  "Another powerful example of the effect of increasing oil prices can be seen in Italy. In 1999, when Italy adopted the euro, the country’s annual trade surplus was $22 billion. Since then, Italy’s trade balance has altered dramatically and the country now has a deficit of $36 billion. Although this shift has many causes, including the rise of imports from China, the increase in oil price was the most important. Despite a decrease in imports of 388,000 barrels per day compared with 1999, Italy now spends about $55 billion a year on imported oil, up from $12 billion in 1999. That difference is close to the current annual trade deficit. The price of oil is likely to have been a large contributor to the euro crisis in southern Europe, where countries are completely dependent on foreign oil."

Note 3. David King is an old acquaintance of mine and for many years we have been working in parallel in surface science studies. I am not sure if there are deep reasons that make people engaged in surface science to move to peak oil studies but, at least, there are at least two cases! 



  1. You know, models are very good, but reality is very stuborn!

    I've always found that models should be used just to describe processes, not the whole evolution of a system, which is much more complicated. This is why in my talks I never present fittings of the world oil production.

    Our problem is that we measure profitability in terms of money instead of energy; you know very well, we should look EROEI, not ROI (or not alone). Otherwise we will never understand why horizons recede. This is the Rainbow pursuit: we run, but we never arrive to the magic cauldron.

    The inelasticity curve (first seen in one of Tom Murpht's posts at Do the Math) has served me to close some Economists' mouths. It seems that this kind of effect appears clearly depicted in their early-course manuals, and it is stated there "to take this as a sign of scarcity". Smiles freeze in their mouths on sighting this graph. So I recommend you (and any other else) to use this graph when discussing with Economists.


    1. @AMT
      What you say about economists is totally true : Economics 101 teaches them that a poorly elastic supply is a sign of scarcity. And then, Economics 102 says that first-necessity goods can be identified thanks to the low elasticity of the demand for them. (In fact, what we observe when scatter-plotting historical data, taking the volume versus the price, is more or less the ratio of these two, which makes the interpretation difficult. If you want to do some respectable work, you will have to consider both separately, which implies using more than the historical record.)

      Nevertheless, Economics 102 gave me a chance to close one more economist's mouth recently. He was telling me that using one's car belongs to first-necessity needs. But in France in the past few years, there was an incentive that was supposed to make fuel-efficient cars cheaper, and fuel-suckers more expensive ; in fact this incentive was turned into some kind of electoral gift, so that you received a governmental subsidy for almost any normal car you bought. The discount was not that high : maybe 2 to 5 percent of the price of the car. But this was enough to boost car sales by approximately 10%. What an elasticity !! This good cannot be of first necessity !! As you describe it, I saw my economist defeated.

      In fact, my "demonstration" is absolutly dishonnest :
      - as I said above, you cannot really infer an elasticity from the historical record.
      - in this example, car buyers do not have a rational behaviour : when they receive a subsidy, they have the feeling that they are paying less for something that is worth more. They would not necessary show the same behaviour if able to buy the same car directly for the lower price.
      - When the incentive is soon to be implemented, people postpone their project of buying the car, and when it comes to an end people run to carsalers in order to benefit from the incentive (this kind of anticipation behavior is a good example of the fact that the historical record alone cannot bring a proof of anything about elasticities).
      - Plus buying a new car is not the same as consuming car-related goods.
      Anyway, my economist was unable to answer, which I think shows how deeply this elasticity-thing is encrypted in their genome.

  2. This comment has been removed by the author.

  3. @AMT

    "you know very well, we should look EROEI, not ROI (or not alone)."

    In fact maybe we could even try to move to some kind of Bancor (as proposed by Keynes prior to Bretton woods) :

    "The Bancor was to be backed by barter and its value expressed in weight of gold. However, this British proposal of introducing a supranational currency could not prevail against the interests of the United States, which at the Bretton Woods conference established the U.S. dollar as world key currency."

    But indexed on the oil barrel or a "bag of raw material" instead of gold.
    Also more or less proposed by China in recent years I think.
    However I'm not an economist at all(god forbid) and not sure feasible or of any help in current context..

  4. A couple of thoughts. We often see a reference to Hubbert's prediction of a US peak in 1970, and a nice bell-shaped curve showing US decline in oil production after that. One difference compared to now, is that as oil extraction in the US became more difficult, oil extraction resources (ie. capital and labour) would have been allocated elsewhere in the world. I imagine this would have the effect of steepening the decline in US production.
    Conversely, the world has no other new resource to which we will direct our oil extraction resources, so perhaps we might expect that the global downside of the peak would (at least initially) be more gradual, because we keep applying all our oil-extraction resources to our existing fields. From this, it's a possibility that, in our determination to maintain (or grow) rates of production, we are bringing forward future production, so that when production rates do start to decrease they do so more rapidly than the US curve would indicate. Such a posiblity is concerning.

    You said :"Over a certain price, production doesn't respond any more," but I don't think this is really supported by Murray and King's scatterplot. I think there could be for a couple of reasons (it's not clear to me which, it could be all three):

    Oil is getting harder to extract, because the "easy" oil is largely gone, and existing fields may be starting to deplete. In this case, the 73 million barrels/day rate just happens to coincide with the depletion of the resource beginning to affect production.
    A subtly different way of viewing it is to imagine that if we kept trying to increase the rate of production (ignoring peak oil for the moment) then incremental production increases would become more and more expensive. I would imagine that a scatterplot in such a scenario would follow an exponential curve (of price vs rate of production).
    Another posibility is what Murray and King diagnose -- a change in the underlying system dynamics that happened to occur at the time that we were extracting about 73 million barrels/day -- I guess this is another way of phrasing idea no. 1.

    In none of these lines of thinking is the 73 million barrels/day particularly significant. One could imagine an alternative universe where we pushed much harder to extract oil (ie. allocated more economic resources sooner), and reached a much higher peak production but sooner. The 73 million barrels/day has just happened to occur at this time. This is all highly speculative, of course!

    Another point of interest that I think should be underlined: As Ugo points out, more expensive oil acts as a drain on the economy because more resources must be applied to extraction as a proportion of GDP, leaving fewer resources to be allocated elsewhere. So, while it is probably true that higher oil prices would lead to an increase in production (though the scatter plot would seem to indicate that _much_ higher prices are now required to make a significant difference), it seems unlikely that the global economy can afford to pay such prices. In other words, there is an economic limit as well.

    Thanks for a great series of posts,

  5. This comment has been removed by the author.

  6. @ Antonio. ROI and EROI are parallel concepts. The vagaries of the financial market may mask the fact that they should be proportional to each other but, in the end, the market is driven by the monetary cost of energy. So, in the post I didn't mention EROEI, only prices, just to keep things simple. Lurking behind, however, is the same concept expressed in energy terms

  7. Gus, I think you are exactly in tune with what I was saying. You are perfectly grasping the concept of the Seneca model, that is we are bringing forward future production. We'll pay that with a more rapid decline and that, indeed, is a "matter of concern", better expressed with the term "ouch!"

  8. What is interesting is that even if people acknowledge peak oil (finally!), they won't acknowledge their life styles.
    Until they have to! (E.g. Greece etc.)

    Ignorance is sweet, reality not.


  9. Dr. Bardi, I think that price is not a big issue. Our financial institutions will find a way to make more money. Governements can 'stimulate' the economy. The ECB can lend billions of euros to private banks at very low interest-rates.

    The big issue is the drop in EROEI.
    When our society has to put more effort in acquiring the same amount of energy (from tarsands, biofuel, wind and solar), the net energy available for the people will decrease. This trend towards a lower EROEI is irreversible.

  10. Oh, well, Hans, in principle you are right. But money is not something that can be just printed in unlimited amounts. It has a relation with the real world and even with EROEI, yes. Economists tend to ignore this relationship, but it does exist. Money is, eventually, nothing but credit. So, when states print money, they create bad credit. But there is no free energy and not even free money. Eventually, reality wins.

  11. In the above, I meant "states print TOO MUCH money"....

  12. When we think of oil, we picture the gas tank analogy. When the needle reaches E for empty is when we are in trouble. The world does in fact have a trillion barrels of oil left to produce. The real analogy is like a Pearl Harbor reconnaissance plane flying its mission over the ocean. The plane flies as far as it can for as high as it can. The pilot fulfils the mission of aerial photography of enemy positions. At a certain point though the pilot knows he must turn around at the HALF WAY point of the gas gauge to make it back home. When the needle reaches at half the tank the pilot MUST RETREAT and DESCEND to make it back to base. When the world has produced as much oil as it ever can in one day (peaked), when it has flown as far as it can for as high as it can the world economy MUST RETREAT and DESCEND.

  13. For sure it will retreat and descend, but maybe more Kamikazee fashion ...

  14. The chart, with its vertical rising price, perfectly shows dynamic of impacting in a wall-like situation: it is ours.