THE PETROLEUM CULTURE
153
recoverable oil. Most of these studies assumed world ultimate recoverable resource volumes of roughly 2 trillion barrels, and that oil
production would peak when 50% of the ultimate resource had been
extracted. In comparison, the United States Geological Survey low
estimate (which they state has a 95% probability of being exceeded) is
2.3 trillion barrels. One analysis fitted the left-hand side of Hubberttype curves to data on actual production while constraining the total
quantity under the curve to two, three, and four trillion barrels for
world ultimate recoverable resource. The resultant peaks were predicted to occur from 2004 to 2030.
Other forecasts for world oil production do not rely on such
curve-fitting techniques to make future projections and/or a priori
assumptions about ultimate recoverable resource. According to the
most recent forecast by the US Energy Information Agency, EIA,
world oil supply in 2025 will exceed the 2001 level by 53% (EIA,
2003). The EIA reviewed five other world oil models and found
that all of them predict that production will increase in the next
two decades to around 100 million barrels per day, substantially
more than the 77 million barrels per day produced in 2001. Several
of these models rely on the new United States Geological Survey
estimates of ultimate recoverable resource for oil. It should be
noted that almost all oil-supply forecasts for which we are able to
examine the predictions against reality had a dismal track record,
regardless of method. Most recent results of curve-fitting methods
showed a consistent tendency to predict a peak within a few years,
and then a decline, no matter when the predictions were made
(Lynch, 2002). It is now a well-established fact that economic and
institutional factors, as well as geology, were responsible for the US
peak in production in 1970 (Kaufmann and Cleveland, 2001), forces
that are explicitly excluded from the curve-fitting models. Thus, the
ability (or the luck) of Hubbert's model and its variants to forecast
production in the 48 lower states accurately cannot necessarily be
extrapolated to other regions. It is too early to tell.
Economic forecasts fare no better in explaining US oil production in the lower 48 states. In the period after the Second World
War, oil production often increased as oil prices decreased, and vice
versa (Kaufmann, 1991), a behavior that is exactly the opposite of
predictions of economic theory, which also assumes that oil prices
will follow an optimal path towards the choke price (i.e. the price at
which demand for oil falls to zero and the market signals a seamless transition to substitutes). In fact, even if such a path exists,
153
recoverable oil. Most of these studies assumed world ultimate recoverable resource volumes of roughly 2 trillion barrels, and that oil
production would peak when 50% of the ultimate resource had been
extracted. In comparison, the United States Geological Survey low
estimate (which they state has a 95% probability of being exceeded) is
2.3 trillion barrels. One analysis fitted the left-hand side of Hubberttype curves to data on actual production while constraining the total
quantity under the curve to two, three, and four trillion barrels for
world ultimate recoverable resource. The resultant peaks were predicted to occur from 2004 to 2030.
Other forecasts for world oil production do not rely on such
curve-fitting techniques to make future projections and/or a priori
assumptions about ultimate recoverable resource. According to the
most recent forecast by the US Energy Information Agency, EIA,
world oil supply in 2025 will exceed the 2001 level by 53% (EIA,
2003). The EIA reviewed five other world oil models and found
that all of them predict that production will increase in the next
two decades to around 100 million barrels per day, substantially
more than the 77 million barrels per day produced in 2001. Several
of these models rely on the new United States Geological Survey
estimates of ultimate recoverable resource for oil. It should be
noted that almost all oil-supply forecasts for which we are able to
examine the predictions against reality had a dismal track record,
regardless of method. Most recent results of curve-fitting methods
showed a consistent tendency to predict a peak within a few years,
and then a decline, no matter when the predictions were made
(Lynch, 2002). It is now a well-established fact that economic and
institutional factors, as well as geology, were responsible for the US
peak in production in 1970 (Kaufmann and Cleveland, 2001), forces
that are explicitly excluded from the curve-fitting models. Thus, the
ability (or the luck) of Hubbert's model and its variants to forecast
production in the 48 lower states accurately cannot necessarily be
extrapolated to other regions. It is too early to tell.
Economic forecasts fare no better in explaining US oil production in the lower 48 states. In the period after the Second World
War, oil production often increased as oil prices decreased, and vice
versa (Kaufmann, 1991), a behavior that is exactly the opposite of
predictions of economic theory, which also assumes that oil prices
will follow an optimal path towards the choke price (i.e. the price at
which demand for oil falls to zero and the market signals a seamless transition to substitutes). In fact, even if such a path exists,