Yergin does not offer a statement of exactly what he means by “peak oil”, though his essay refers to it as a “fear” and a “specter”. Let me therefore begin my remarks with a clarification of exactly what I intend to discuss. I propose the following three propositions as the core claims that need to be evaluated:

1. The annual flow rate of oil production from a given reservoir eventually reaches a maximum, after which it declines.
2. The annual flow rate of total global oil production will eventually have to decrease as a necessary consequence of (1).
3. This peak in global production will be reached relatively soon.

Of these statements, I honestly don’t understand how a reasonable person could dispute (1). You could almost take it as tautological, and furthermore point to many, many examples of fields that passed their peak production long ago. I likewise see neither a conceptual nor an empirical basis for challenging (2). Thus it seems to me that the relevant debate is whether proposition (3) has any merit, and exactly what one means by “soon.” That question may or may not be what Yergin was intending to address with his essay. But since for me it is the core question, I would like to comment here on the implications of what Yergin wrote for what I perceive to be the main question of interest.

Yergin wrote:

This is actually the fifth time in modern history that we’ve seen widespread fear that the world was running out of oil. The first was in the 1880s, when production was concentrated in Pennsylvania and it was said that no oil would be found west of the Mississippi. Then oil was found in Texas and Oklahoma. Similar fears emerged after the two world wars. And in the 1970s, it was said that the world was going to fall off the “oil mountain.” But since 1978, world oil output has increased by 30%.

Let me begin by interjecting a few facts into this discussion. Annual production from the state of Pennsylvania in fact did peak in 1891, when the state produced 31.4 million barrels. Last year, the state only produced 3.5 million barrels, which is 89% below its historical peak. Texas production peaked in 1972, and is now 67% below the historical peak. Oklahoma peaked in 1927, and is now 75% below those levels.

Back to Yergin:

The date of the predicted [global] peak has moved over the years. It was once supposed to arrive by Thanksgiving 2005. Then the “unbridgeable supply demand gap” was expected “after 2007.” Then it was to arrive in 2011. Now “there is a significant risk of a peak before 2020.”

Again I’d like to bring up a few facts. Although it is true that global production did not fall between 2005 and 2010, it is also accurate to observe that it did not grow very much, rising only 2.2 million barrels/day (which represents 2.6% of 2005 levels) over these 5 years. Over these same 5 years, China increased its consumption by 2.5 mb/d. Thus, although the world did produce more, everybody in the world outside of China had to make do with less.

Total global oil production, in millions of barrels per day, annual 2002-2010 (data source: EIA).

Even in the absence of these facts, there’s a real problem with Yergin’s line of argument for the question at hand, and it troubles me because I have seen the same argument raised almost every time someone takes the skeptic’s position on the question of peak oil. Suppose I was trying to convince you that you are a mortal being, and your counterargument was, “but that’s what you said in 2005, and I didn’t die then! You said it again in 2007 and 2009, and each time you were wrong. Why should I believe you this time?”

Perhaps acknowledging one’s own mortality is a similar proposition to embracing the possibility that global oil production need not continue to rise forever.

In any case, I was not among those who claimed that the peak would arrive by Thanksgiving 2005, nor 2007, nor 2011. But I am among those who did claim, and still believe, that the slow rate of increase in annual oil production over the last 5 years has caused significant economic problems for countries like the United States.

Moreover, if having been wrong in the past were a valid reason to disregard everything someone says, it might be wise to ponder these words that Daniel Yergin wrote in 2005:

There will be a large, unprecedented buildup of oil supply in the next few years. Between 2004 and 2010, capacity to produce oil (not actual production) could grow by 16 million barrels a day — from 85 million barrels per day to 101 million barrels a day — a 20 percent increase. Such growth over the next few years would relieve the current pressure on supply and demand.

Dissecting what went wrong with that prediction is a topic for another occasion. I believe it was based on a careful, thoughtful analysis, and provides an interesting case study in some of the challenges facing anybody who tries to make these kinds of predictions. But I do feel that the meme of “don’t listen to the peak oil nuts, because they’ve always been wrong before” should have gotten a bit tiresome at this point.

The other main issue Yergin takes up in his most recent contribution is concerns about the logistic curve that M. King Hubbert successfully used in 1956 to predict the peak of U.S. oil production around 1970. Here again I feel that Yergin is attacking a straw man. The core issue should be the validity of propositions 1-3 above. While it is true that some people have used variants of Hubbert’s approach to propose particular dates for when “soon” might actually be, I do not include myself among them. To cast those mathematics as if they were the central issue in the debate is doing a significant disservice.

I submit that meeting the growing global demand for crude oil over the last five years has posed significant challenges for the world economy. And those who worry that the next 5-10 years might be like the last should not be dismissed as crackpots.