Introduction Physical Dependence Physical Vulnerability Economic Measures ConclusionMeasures of Oil Import Dependence
by
James M. Kendell
Measures of oil import dependence or vulnerability can be divided into physical and economic dimensions. Physical measures of dependence have been used most frequently to assess the level of U.S. needs for imported oil. However, measures of import vulnerability—whether physical or economic—are likely to be more useful than measures of dependence in assessing U.S. energy security. |
Introduction
Physical Dependence
Physical Vulnerability
Economic Measures
Conclusion
In 1977 the United States imported a record 46.5 percent of the oil it needed to fuel its vehicles, heat its homes, and run its industry. In reaction to rising prices and such high levels of imports, the Nation established a Department of Energy, spent billions of dollars on researching and finding new sources of energy supply, and redesigned its cars, houses, and factories to make them more energy efficient. Yet last year, when the United States broke its 20-year record for oil import dependence, few voices were heard noting, let alone decrying, the high levels of imports. In the interim, analysts and policymakers had learned that simple measures of physical dependence do not tell the whole story of oil imports.
This paper explores the meaning and value such measures as “net imports as a percentage of product supplied” when used as indicators of energy security. While the limits of this particular dependence measure are now generally understood, policymakers still need good measures of energy security to tell them (and the voters) when U.S. vulnerability is growing as a result of increased oil imports. In choosing oil security measures, one of the most important distinctions is between oil import dependence and oil import vulnerability. Knowing that the Nation imports 2 percent or 50 percent of its oil tells how dependent it is, but not how vulnerable it is to oil price shocks and to oil supply disruptions. The distinction between dependence and vulnerability has been made for years by oil analysts, but with the United States poised to move beyond 50 percent dependence, it is worth drawing the distinction once again.
A variety of measures have been used over the years by the Energy Information Administration (EIA) to gauge the significance of oil imports (Table 1). In addition, a comprehensive series of measures was used by the U.S. General Accounting Office (GAO) in a recent study of energy security. The GAO measures are notable for their attempt to cover the range of physical and economic aspects of both import dependence and vulnerability. The GAO measures were developed in consultation with EIA and other offices in the Department of Energy.
Table 1. Oil Import Dependence and Vulnerability Measures
Measure |
Physical Dimension |
Economic Dimension |
Dependence |
Import Share of Product Supplied |
Value of
Imports |
Vulnerability |
Percent
of World Exports |
Consumption
per Dollar of GDP |
The EIA has regularly published a measure of oil import dependence since 1979. In March 1979, gross oil imports as a percentage of product supplied began to appear as an ongoing graph in the Monthly Energy Review. The gross import dependence percentage had risen steadily from 1967 through 1977 and appeared ready to break the 50 percent barrier (Figure 1). Consequently, the gross percentage was published in EIA’s leading publication, the Monthly Energy Review. (At the time, gross imports were labeled ”direct” imports; today EIA uses the term ”total” imports for gross imports.)
Figure 1.
Petroleum Imports as a Percentage of
Products Supplied, 1970-2020
Sources: History:
Energy Information
Administration (EIA), Annual Energy
Review 1996, DOE/EIA-0384(96)
(Washington, DC, July 1997).
Projections: EIA, Annual Energy
Outlook 1998, DOEEIA-0383(98)
(Washington, DC, December 1997).
The trouble with using gross imports in the numerator of this measure is that it overstates U.S. dependence on imported oil. For, if all else were equal, rising exports would mean a higher percentage of gross oil import dependence. In the 1970s, when oil exports were relatively small, the distinction between gross and net imports was minor. Exports averaged only 267,000 barrels per day, with Canada, Japan, and Mexico as major recipients of U.S. oil. But by 1982 exports had more than tripled to 815,000 barrels per day, with the Virgin Islands, Puerto Rico, the Netherlands, and Canada as major export partners. Thus, in December 1982, the Monthly Energy Review began to publish net imports as a percentage of product supplied.
Even though the EIA stopped publishing gross imports as a percentage of products supplied in 1982, the American Petroleum Institute (API) continued to use this measure. API argued that there was no market for the petroleum products exported by the United States, and that the exports were irrelevant to a dependence calculation. The counter argument is that some U.S. exports could easily be consumed domestically. Indeed, more than half of U.S. oil exports in 1997 were crude oil, natural gas liquids, gasoline, gasoline blending components, jet fuel, and distillate fuel oil—all marketable liquids within the United States.
In 1993, gross oil imports reached 50 percent of product supplied, exceeding the previous record set in 1977 and even prompting calls for legislative action. Since then gross oil imports have grown, and they are expected to continue growing over time.
In August 1995 the EIA made its most recent change to the publication of oil import dependence numbers, by adding ”gross imports as a percentage of product supplied” alongside its publication of ”net imports as a percentage of product supplied” in the Monthly Energy Review. An accompanying note stated that: ”EIA believes that the net-imports definition gives a clearer indication of the fraction of oil consumed that could not have been supplied from domestic sources and is thus the most appropriate measure.”1
As noted above, in 1997 the United States exceeded its 20-year record for net oil imports as a percentage of product supplied.2 The Nation imported almost 48 percent of its net petroleum supply in 1997, compared with the previous record of 46.5 percent in 1977. The Annual Energy Outlook 1998 (AEO98) reference case projects that net dependence will exceed 50 percent in 2000 and rise to 66 percent in 2020.3
The AEO began to publish an oil import dependence measure in 1996. As might be expected, the measure chosen was net imports as a percentage of product supplied. Until then, AEO users had to calculate their own measures, which led to some interesting questions. More than once, EIA was called to explain why it was showing import dependence of over 70 percent in 2015. It is possible to reach such high numbers, but only by doing the calculation on a gross basis with British thermal units (Btu) rather than barrels. The argument for doing a net, rather than gross, calculation is stated above. The argument for using barrels, rather than Btu, is that a physical measure of dependence (or vulnerability) ought to use physical units, rather than a heat value. After all, producers and consumers typically sell and buy oil in barrels, not in Btu.
When the EIA began publishing measures of oil import dependence in 1979, the Monthly Energy Review also began publishing a graph of dependence on oil imports specifically from the Organization of Petroleum Exporting Countries (OPEC), because OPEC was widely viewed as controlling the world oil price. The graph showed gross imports from OPEC as a percentage of U.S. product supplied (although gross and net imports in this case are virtually the same). A few years later, in 1982, the Monthly Energy Review also began publishing a graph showing dependence on Arab OPEC oil imports, because the Arab members of OPEC had stopped exporting oil to the United States in 1973-1974. In 1982 the calculations for OPEC and Arab OPEC were changed to net imports as a percentage of U.S. product supplied, just as the total dependence calculation had been switched to net imports. Both the OPEC and Arab OPEC net percentage peaked in 1977, the same year of the overall peak.
In 1995, in the wake of the Iran-Iraq war, the Persian Gulf war, and strategic thinking about the vulnerability of oil exports from the Persian Gulf, the Monthly Energy Review stopped publishing OPEC and Arab OPEC percentages and switched to imports from the Persian Gulf region as a percentage of product supplied.4 Like the other measures, that percentage peaked at more than 13 percent in 1977. The 1977 peak is not projected to be exceeded until 2017, according to the AEO98 (Figure 2). By 2020 the Persian Gulf is expected to supply 14 percent of U.S. consumption.
Figure 2. Net
Petroleum Imports as a Percentage
of Products Supplied, 1973-2020
Other variations of such dependence measures have also been used or suggested. One refinement could be to add or subtract the net oil stock change to net petroleum imports in the numerator, on the theory that imports to build stocks are not being consumed.5 On an annual basis, such a refinement would make little difference, but it could make a substantial difference on a monthly basis.
Another, more significant variation would exclude Canada and Mexico from the numerator (Figure 2).6 Canadian and Mexican supplies are closer to the United States and are, in theory, less vulnerable to interruption than supplies that must travel long distances on the open sea. By this measure, dependence on other imports does not reach the 1977 peak until 2012.
While measures of oil import dependence may be of some interest, they offer a limited guide to energy security. At 48 percent net dependence on imported oil this year, is the United States any less secure than it was at 42 percent in 1990, 43 percent in 1979, or 35 percent in 1974? Other countries, including Japan, Sweden, and Spain, have managed to grow and prosper despite almost complete dependence on imported oil.
Oil dependence does not necessarily mean that the United States is vulnerable to an oil disruption. If the world oil supply came from many small producers and one of them suddenly stopped exporting oil, it would have little effect on U.S. and world supplies and prices, even at a high rate of U.S. dependence. The distinction between dependence and vulnerability suggests that concentration is a key factor in the security of our oil supply.
Concentration of world oil production in the Persian Gulf is one of the measures used by the GAO in its December 1996 report, Energy Security: Evaluating U.S. Vulnerability to Oil Supply Disruptions and Options for Mitigating Their Effects.7 The Persian Gulf percentage of world oil production, which declined from 1976 to 1985, has been generally rising since then and is expected to continue rising through 2020, according to the AEO98 (Figure 3).
Figure 3.
Persian Gulf Share of World Oil
Production and Exports, 1975-2020
Sources: History:
Energy Information Administration (EIA),
International Petroleum Statistics Report, DOE/EIA-0520(97/07)
(Washington, DC, July 1997); and EIA, Monthly Energy Review,
DOE/EIA-0035(98/02) (Washington, DC, February 1998).
Projections: EIA, Office of Integrated Analysis and Forecasting.
Perhaps more important than the Persian Gulf share of world oil production is its share of the world export market: if most Persian Gulf oil production were consumed in the Persian Gulf, a supply disruption would not directly affect U.S. vulnerability. In fact, however, if those exports were cut off, the effect would be immediate and direct, because a disruption in one part of the world quickly affects supplies and prices in the rest of the world. The peak for Persian Gulf oil exports as a percentage of world oil exports was in 1974, when they accounted for more than two-thirds of the oil traded in world markets. The Persian Gulf share of world oil exports has risen since the oil price collapse of the mid-1980s, but it is not expected to surpass the 1974 level until after 2020. A graph of ”Persian Gulf Share of Worldwide Oil Exports” appeared in the AEO beginning with AEO97.
In 1977 two of the top five U.S. oil suppliers were Persian Gulf countries (Iran and Saudi Arabia), and two more were also in the Eastern Hemisphere (Libya and Nigeria). But 20 years later, only two of those Eastern Hemisphere suppliers, Saudi Arabia and Nigeria, remained in the top five. In 1997, three of the top U.S. oil suppliers were in the Western Hemisphere (Venezuela, Canada, and Mexico).
Even as the world and the United States have moved away from dependence on Persian Gulf oil, however, the reliance on large suppliers has increased. In 1977 the top five U.S. oil trading partners provided the United States with the equivalent of 25 percent of product supplied (on a gross basis). In 1997 the top five provided the United States with 36 percent of product supplied. Over these 20 years gross import dependence has increased by 5 percentage points, whereas dependence on the top five suppliers has grown by 11 percentage points.
Emergency, noncommercial inventories are one strategy to cope with supply disruptions. In 1974 the United States experienced its most significant supply interruption with the cutoff of about 18 million barrels per day, or about 55 percent of the world export market. During that disruption, the world oil price tripled, from about $4 a barrel to about $12 a barrel. In 1990, the Iraqi invasion of Kuwait meant the loss of 4.3 million barrels of oil production per day, or about 13 percent of the world export market. This led to a doubling in the world oil price from July to October 1990, from about $16.50 to about $33 a barrel.
By 1990 the United States and other governments had created emergency stockpiles of oil as a buffer against disruption. The invasion of Kuwait showed that the United States and other governments were willing to use their stockpiles. A noncommercial measure, “Days of Net Petroleum Imports in the Strategic Petroleum Reserve,” is published in the Annual Energy Review. It shows that the U.S. Strategic Petroleum Reserve (SPR) peaked at 115 days of supply in 1985 and has now declined to 63 days. Assuming that the SPR does not expand or contract, coverage will decline to 35 days in 2020 as consumption grows.
Combining noncommercial and commercial stocks provides a somewhat broader measure of the ability of inventories to respond to supply disruptions. Since 1985, available commercial stocks in the Organization for Economic Cooperation and Development (OECD) countries have fluctuated between 25 and 30 days of supply. Assuming that commercial pressures keep stockpiles from expanding, while consumption continues to grow, the supply would slip to 20 days in 2020.8
Besides stockpiles, surge capacity or excess world production capacity is another source of supply. Historically, excess capacity has responded primarily to prices, building up during periods of high prices and declining during periods of low prices. A buildup occurs during a high-price period such as the early 1980s, as consumers conserve and producers rush to find more oil and cash in on high prices. If oil prices remain at their current moderate levels through 2020, excess capacity can be expected to decline from 3.4 to 2.4 million barrels per day in 2013, before rising to 3.2 million barrels per day in 20209 (Figure 4).
Figure 4.
World Excess Oil Production
Capacity, 1975-2020
Source: Energy Information Administration,
Office of Integrated Analysis and Forecasting.
From 1992 to 1994, EIA combined some of these vulnerability indexes to create a composite “Index of OPEC Dependence.” At first named the “Vulnerability Index,” this composite measure was published in the International Energy Outlook.10 The composite index was for three measures: the percentage of world oil demand supplied by OPEC, OECD oil stock levels, and excess OPEC crude oil production capacity. The index was a weighted average of 50 percent for excess capacity, 30 percent for OPEC market share, and 20 percent for available stocks. Each of the variables was given a weight of 100 when at the greatest dependence and 0 at the least dependence. The advantage of combining the three measures is that it allows the measure of OPEC market power to be mitigated by measures of stocks and excess capacity.
In 1994 the index showed that dependence on OPEC was expected to increase and, by 2010, to be close to the levels of the early 1970s. If the same index were calculated with AEO98 projections, the high 1973-74 levels probably would not be reached until after 2020, primarily because of the significant increases in non-OPEC oil production that have been projected since 1994.
In addition to supply-side measures of oil vulnerability, demand-side measures have been constructed. In the transportation sector, the GAO study used ”Oil as a Percentage of Total Energy Used in Transportation.” Dependence on oil in the transportation sector is projected to decline from about 97 percent at the present time to about 95 percent in 2020, as alternative fuel consumption grows, with compressed natural gas leading the way. Another demand-side measure is oil consumption per capita, a measure that the EIA has occasionally published. By this measure, Americans consumed 26 barrels of oil each in 1970 and again in 1997, with a projected increase to 28 barrels each in 2020.
None of these simple vulnerability measures assesses the probability of disruption. An ideal measure might include a disruption probability, based on the level of concentration of control of the world export market and on other economic, political, and military factors.
Economic measures of oil security are at least as important, if not more so, than physical measures. Physical shortages or disruptions are quickly manifested as price increases. Thus, any discussion of disruption quickly leads to discussion of prices. Since September 1991, the Monthly Energy Review has reported the value of petroleum imports and exports; long before that, in December 1978, it began reporting the value of total energy imports and exports.
The AEO97 began to report gross expenditures on imported oil as a measure of the impact of oil imports. In AEO98 the measure was refined to net expenditures on imported oil. This calculation was added to the AEO as a result of an article coauthored by then-Deputy Secretary of Energy Charles Curtis.11 It cited the steady rise to an annual $100-billion-plus projected cost for imports as an argument for increased spending on energy research and development. As startling as the doubling of the value of oil imports may be, even by 2020 the total value is not expected to exceed the 1980 peak of $138 billion in constant 1996 dollars (Figure 5). Oil imports were very expensive in 1980 because the price of oil was nearly $62 a barrel. By 1986 the price of oil had dropped by two-thirds and the number of barrels of net oil imports had declined by 15 percent, causing the value of imported oil to decline to about $43 billion in 1996 dollars. As U.S. oil consumption increases and production declines, the cost of imported oil is expected to rise through 2020.
Figure
5. Net Expenditures for Imported Crude Oil
and Petroleum Products, 1975-2020
Sources: History:
Energy Information Administration (EIA),
Monthly Energy Review, DOE/EIA-0035(98/02)
(Washington, DC, February 1998). Projections: EIA, Office of
Integrated Analysis and Forecasting.
While it is interesting to know that Americans pay a rising bill for imported oil, in the context of total imports of goods and services, the bill for imported oil is rather small. From a peak of 39 percent in 1980, net oil spending as a percentage of total imports fell to a mere 6 percent in 1995 and is expected to fall below 3 percent in 2020 (Figure 6). The most significant decline in the oil percentage of imports occurred from 1980 to 1986, from 39 to 7 percent, as the world oil price dropped to less than one-third of its 1980 value, physical barrels of U.S. net oil imports declined by 15 percent, and total U.S. imports increased. During the same period OPEC lost its grip on the world oil price, as high world oil prices pushed down demand and pulled up non-OPEC oil production by 15 percent. Aided by a strong dollar, total U.S. imports of goods and services, particularly automobiles and capital goods, grew by 64 percent from 1980 to 1986, even as the value of oil imports was declining. In the future, oil imports and oil prices are expected to grow, but not as fast as total imports.
Figure
6. Oil Expenditures as a Percentage of Total
Spending on Imported Goods and Services, 1975-2020
Measures such as the value of oil imports and the oil percentage of total expenditures on imported goods and services might be thought of as economic measures of oil dependence. As with physical measures, economic dependence measures are probably less valuable to long-run thinking about energy security than are economic vulnerability measures.
Oil intensity—oil consumption per dollar of gross domestic product (GDP)—is one measure of the economy’s vulnerability to oil disruptions. As oil intensity declines, an oil disruption of a given size will have less effect on the economy. For example, if reduced oil intensity comes about through increased mileage per gallon, a disruption should also have less effect on drivers. However, oil intensity might also decline as vehicle miles traveled decline because of an economic slowdown.
The EIA begin to publish a measure of energy intensity in the Monthly Energy Review in March 1979 and continues to do so. A measure of oil intensity has never been published on a regular basis, however. (Petroleum intensity was included in a graph in the EIA report, Energy Conservation Indicators 1986.12) The GAO chose oil intensity as one of its measures of vulnerability to supply disruptions.Since 1970 oil intensity has generally declined, even though oil consumption has increased, because GDP has increased even faster (Figure 7). The United States experienced a significant decline in oil intensity from 1976 to 1985, as high oil prices squeezed out the most expensive uses for oil and the easiest to switch to other fuels, such as home heating and electricity generation. Oil intensity is expected to continue declining in the future, falling by 16 percent from 1996 through 2020.
Figure 7. U.S. Oil Intensity, 1970-2020
A similar measure that has occasionally been used in EIA publications is energy expenditures per dollar of GDP. Petroleum expenditures per dollar of GDP have also been used. These measures show that end-use energy and petroleum expenditures peaked in 1980-1981 at 14 and 9 percent of GDP, respectively. After that, the energy and petroleum shares declined steadily to about 7 and 3 percent today.
Measures of oil dependence and vulnerability can be divided into physical and economic dimensions. Physical measures describe the relative level of imports or the prospects for shortages and disruptions. Economic measures are less familiar. They describe the cost of imports or the prospects for price shocks.
Whether physical or economic, in the long run measures of vulnerability are likely to be more useful to policymakers than measures of dependence. Measures of dependence simply show the extent of the Nation’s imports. By themselves they provide little information about energy security. In contrast, measures of vulnerability show the meaning of the imports, indicating the Nation’s vulnerability to shortages, disruptions, and price spikes. By studying measures of vulnerability, policymakers can gauge their progress toward insulating the Nation from the harmful effects of sharp changes in the world oil market.
To:
Motor Fuels Tax Trends and Assumptions
Back To:
Issues in Midterm Analysis and Forecasting 1998 Homepage
File last modified: July 22, 1998
Contact the National
Energy Information Center for questions at (202) 586-8800
If you having technical problems with this site, please
contact the EIA Webmaster at wmaster@eia.doe.gov