The author is group senior vice president, Corporate Strategies, Eni Spa, Rome, Italy.


After World War I, the United States was shaken by predictions of the exhaustion of domestic oil. Even the head of the U.S. Geological Survey (USGS)— among many others—deliv-ered a verdict of gloom in 1919: The country would run out of oil within 9 years! (1)

Facing mounting hysteria, President Coolidge set up the Federal Oil Conservation Board in 1924, to draft legis-lation to preserve national re-sources.

After the conversion of Great Britain’s naval fleet from coal to oil in 1914, the UK also feared that it would be vulnerable to oil shortages and moved to secure its grip on the Persian Gulf. These cycles of hysteria followed by new bonan-zas have continued to the present. Thus, it is not surprising that a new wave of “oil doomsters” predicting imminent petroleum scarcity has gained momentum (2–4).

The worst effect of this recurring oil panic is that it has driven Western political circles toward oil imperialism and attempts to assert direct or indirect control over oil-producing regions. Yet the world is not running out of oil, and catastrophic views fail to take into account the complex reali-ty that will allow reliance on abundant sup-plies for years to come.

The current model of oil doomsters is derived from K. M. Hubbert (5). The model is conceptually simple, but based on several assumptions. The first is that the geological structure of our planet is well known and thoroughly explored, so that discovery of unknown oil fields is highly improbable.

Second, to resolve problems connected with erratic distribution and production from thousands of oil fields and uncertainty of future discoveries, production is assumed to follow the “Central Limit Theorem” from statistics. This theorem states that the sum of a large number of erratic variables tends to follow a normal distribution and assumes a bell-shaped curve (see figure above).

Starting from zero, production grows over time until it peaks when half of the re-coverable resources have been extracted (“midpoint depletion”). Then, production irreversibly declines at the same rate at which it grew. The area under the curve shows the cumulative production of an oil field or the “ultimate recoverable re- sources” (URR) it holds and their life-span.

Accordingly, to forecast Earth’s URR, one needs to process worldwide production and discovery trends and geological data.

In 1956, Hubbert accurately predicted the peak oil production point of the U.S. lower 48 states.

The Hubbert curves do not delineate the complex and dynamic nature of oil produc-tion and reserves in the world, because they are the product of a static model that puts an unjustifiable faith in geology and does not consider technology and cost/price functions. The model’s success in predict-ing U.S. peak production merely reflected the peculiar nature of this area, which is the most intensively explored and exploited in the world. Elsewhere, the pattern of pro-duction is not rendered by a bell curve but is marked by large discontinuities (see fig-ure on next page).

Using different versions of the Hubbert model, several geologists have made pre-dictions in the last 20 years of an imminent crisis in oil availability that subsequently had to be revised. The most eminent among them is C. Campbell, who predicted that 1989 was the year of “peak” production (6). The estimates have been increasing steadily (see table, next page).

Before looking at the real-world situa-tion in more depth, it is necessary to clear up some points, beginning with the distinction between “resource” and “reserve.”

The former indicates the overall stock of a mineral in physical terms, without any associated economic value and/or estimation of its likeli-hood of being extracted. In other words, there may be large quantities that can nev-er be used because of the high cost or the impossibility of recovery, as in the case of the gold dispersed in the oceans. The concept of “re-serves”— like that of “recov-erable resources”—involves an economic assessment of the possibility of producing a part of the overall resources. In the oil sec-tor, there are additional definitions—the most important being that of “proven re-serves,” which include only those that can be economically produced and marketed at the present time according to existing tech-nologies and demand. Nearly all of the es-timates of the world’s oil URR, including those by oil doomsters, do not take into ac-count the so-called “nonconventional oils”—such as Canadian tar-sands and Venezuelan and Russian heavy oils—even though the availability of these resources is huge and the costs of extraction falling.

Although hydrocarbon resources are ir-refutably finite, no one knows just how fi-nite.

Oil is trapped in porous subsurface rocks, which makes it difficult to estimate how much oil there is and how much can be effectively extracted. Some areas are still relatively unexplored or have been poorly analyzed. Moreover, knowledge of in-ground oil resources increases dramati-cally as an oil reservoir is exploited.

For example, the Kern River field was discovered in California in 1899.

Calculations in 1942 suggested that 54 million barrels remained. However, in 1942 “…after [43] years of depletion, ‘re-maining’ reserves were 54 million barrels.

But in the next [44] years, it produced not 54 but 736 million barrels, and it had an-other 970 million barrels ‘remaining’ in 1986. The field had not changed, but knowledge had….” (7). This is but one of hundreds of cases reported in oil-related literature that underscore the inherently dy-namic nature of oil reserves. As Klett and Schmoker have recently demonstrated, from 1981 to 1996 the estimated volume of oil in 186 well-known giant fields in the world [>0.5 billion (10 9 ) barrels (Bbl) of oil, discovered before 1981] increased from 617 to 777 Bbl without new discover-ies (8). Indeed, many studies have proved the phenomenon of “reserve growth”—i.e., that “additions to proven recoverable vol-umes are usually greater than subtractions” (8). This occurs because of four fundamen-tal elements: technology, price, political decisions, and better knowledge of existing fields—the last of these being possible on-ly through effective and intensive drilling.

We anticipate that this trend will contin-ue.

Consider, for example, the most recently discovered oil frontier in the world, Kazakhstan, and its major finding—the gi-gantic Kashagan field. Geological estimates about the general area around Kashagan (the Kazakh North Caspian Sea Shelf) have ex-isted for decades, but they only indicated the possibility of hydrocarbon deposits. After the first advanced geological appraisal was conducted by international oil companies in the second half of the 1990s, the area was deemed to hold between 2 and 4 Bbl. In 2002, after completion of only two explo-ration and two appraisal wells in the Kashagan field, estimates were officially raised to 7 to 9 Bbl of producible reserves.

In February 2004, after four more explo-ration wells in the area, they were raised again to 13 Bbl. This is only the beginning, because this area spans over 5500 sq km, and six exploration wells are a modest indi-cator of future potential. Moreover, there are many other oil fields yet to be explored in this area (including Kairan, Aktote, and Kalamkas), that have a geological structure similar to that of Kashagan.

Thanks to new exploration, drilling, and recovery technology, the worldwide finding and development cost per barrel of oil equivalent (boe) has dramatically declined over the last 20 years, from an average of about $21 in 1979–81 to under $6 in 1997–99 (in 2001 dollars) (9). At the same time, the recovery rate from world oil fields has increased from about 22% in 1980 to 35% to-day.

All these factors partly ex-plain why the life-index of world reserves (gauged as the ratio between proven oil re-serves and current production) has constantly improved, pass- ing from 20 years in 1948 to 35 years in 1972 and reaching about 40 years in 2003.

Today, all major sources estimate that proven world oil reserves exceed 1 trillion (10 12 ) barrels, while yearly consumption is about 28 billion barrels (10–13).

Overall, the world retains more than 3 tril-lion barrels of recoverable oil resources (14).

Critics could note that new oil discov-eries are only replacing one-fourth of what the world consumes every year (fol-lowing a declining trend that began in the mid- 960s), and that increases in re-serves largely derive from upward revi-sions of existing stock. However, the real issue is that neither major producing countries nor publicly traded oil compa-nies are keen to invest money in substan-tial exploration campaigns. The countries richest in oil have minimized their oil in- estments during the last 20 years, main-ly for fear of creating a permanent excess capacity such as that which provoked the crisis in 1986 (when oil prices plummet-ed to below $10/bbl). In fact, countries such as Saudi Arabia or Iraq (which to-gether hold about 35% of the world’s proven reserves of oil) produce petroleum only from a few old fields, although they have discovered but not developed more than 50 new fields each. Moreover, in countries closed to foreign investments, the technologies and techniques used are, in most cases, obsolete.

Nevertheless, international public oil companies have faced two sets of limits to their expan-sion in the last 20 years. The first is inaccessibility to foreign in- estment in the largest and cheap-est reserves—those in the Persian Gulf. Second are the demands of financial markets, which for years have insisted that compa-nies provide unrealistic, short-term financial returns that are in-consistent with the long-term na-ture of oil investments. This has compelled private operators to reject op-portunities that would normally be deemed economically worthwhile. This financial pressure partly explains recent proven re-serve downgrading by some oil companies, starting with the amazing cuts announced by the “supergiant” Shell Group (15).

Indeed, this Anglo-Dutch oil company has not lost its resources. This picture has noth-ing to do with physical scarcity of oil.

The Age of Coal began when declining supplies of wood in Great Britain caused its price to climb. Two centuries later, oil took the place of coal as “the king of ener-gy sources” because of its convenience and its high flexibility in many applications, but coal was neither exhausted nor scarce.

Oil substitution is simply a matter of cost and public needs, not of scarcity. To cry wolf ” over the availability of oil has the sole effect of perpetuating a misguided ob-session with oil security and control that is already rooted in Western public opinion— an obsession that historically has invariably led to bad political decisions.

Years Production (Bbl/ year)

1850 1875 1900 1925 1950 1975 2000 2025 2050

Peak production (or “midpoint depletion”

Cumulative production or ultimate recoverable resources (URR)

The Hubbert curve (United States). Bbl, billion (10 9 ) barrels. 21 MAY 2004 VOL 304 SCIENCE www.sciencemag.org1115

References and Notes

1. D. Yergin, The Prize: The Epic Quest for Oil, Money, and Power (Simon & Schuster, New York, 1991), p.194.

2. “The end of the oil age,” The Economist, 23 October 2003, pp. 11, 61–63.

3. D. Goodstein, Out of Gas—The End of the Age of Oil (Norton, New York, 2004).

4. Deutsche Bank, “Hubbert’s pique,” Global Energy Wire,June 2003.

5. K. M. Hubbert, “Nuclear energy and the fossil fuels,” in Drilling and Production Practice series (American Petroleum Institute, Washington, DC, 1956).

6. C. Campbell, Oil Price Leap in the Early Nineties (Noroil, Kingston-upon-Thames, UK, 1989).

7. M. Adelman, The Genie Out of the Bottle (MIT Press, Cambridge, MA, 1995).

8. T. R. Klett, J. W. Schmoker, AAPG Memoir No. 78, 107 (2003).

9. International Energy Agency, World Energy Outlook 2001 Insights(Organization for Economic Cooperation and Development/IEA, Paris, France, 2001).

10. Oil Gas J. (December 2002).

11. Eni—World Oil and Gas Review (May 2003).

12. BP’s Statistical Review of World Energy 2003 (British Petroleum, London, June 2003).

13. World Oil (August 2003).

14. USGS, World Petroleum Assessment 2000 (USGS, Washington, DC, 2000).

15. PIW ( Petrol. Intell. Wkly.), 19 January 2004.

16. M.A.Adelman, M. C. Lynch, Natural Gas Supply to 2100 (International Gas Union, oersholm, Denmark, 2002).

Years Production (Bbl/ year)

1960 1967 1972 1977 1982 1987 1992 1997 2002

Historical behavior of oil production in Egypt (16).


Petroleum URR (Bbl) (year) Hubbert Campbell USGS

1350 (1969) 1578 (1989) 1796 (1987)

2000 (1973) 1650 (1990) 2079 (1991)

1750 (1995) 2272 (1994)

1800 (1996) 3021 (2000)

1950 (2002)