International Energy Outlook 1999 - World Energy Consumption

The IEO99 projections indicate substantial growth in world energy use,including substantial increases for the developing economies of Asia and South America. Resource availability is not expected to limit the growth of energy markets.

In 1998, expectations for economic growth and energy market performance in many areas of the world were dashed. The Asian economic crisis proved to be deeper and more persistent than originally anticipated, and the threat and reality of spillover effects grew through the year. Oil prices crashed. Russia’s economy collapsed. Economic and social problems intensified in energy- exporting countries and in emerging economies of Asia and South America. Deepening recession in Japan made recovery more difficult in Asia and increased the prospects for slower economic growth in Europe and North America. The rate of worldwide economic expansion in 1998 fell by a third relative to that achieved through most of the 1990s.

Growth in energy demand, the hallmark of recent market performance, has been severely constrained by current economic conditions. Shortage has been replaced by surplus in the hardest hit countries. Fear of spillover effects to other regions, as well as recession in Japan and economic collapse in Russia, caused a range of capital- intensive infrastructure development projects involving power generation, pipeline transport, and liquefied natural gas (LNG) supply to be scaled back or put on hold. Uncertain financial markets have made it more difficult to gain financial backing for some projects. Exploration and development expenditures for oil and gas production were in sharp decline in most parts of the world at the close of 1998. Russia’s devaluation and apparent default on international debt were choking off investments designed to maintain or expand its role in international oil and gas markets. Given these developments, the near-term prospects for energy markets have become more risky and uncertain; and to the extent that energy investment is affected, so too are intermediate and long-term prospects for energy supply and demand.

In light of these developments, the International Energy Outlook 1999 (IEO99) contains revisions to long-term projections of growth in energy supply and demand trends, relative to those presented in IEO98. The revisions are relatively modest, however: a near-term pause in energy demand growth; somewhat (6 percent) lower levels of aggregate demand in the later years of the projection; and lower projected oil prices for the next few years, followed by recovery to last year’s baseline by 2007.

The key themes associated with the IEO98 reference case projections remain: substantial increases in energy demand, based mostly on fossil fuels are in prospect; a major component of growth in world energy demand will be accounted for by developing economies of Asia and South America; natural gas is increasingly the fuel of choice for future electric power generation; oil demand growth is an increasing function of trends in transportation activity; and, over the projection period of about two decades, resource availability does not limit the development of energy markets.

Each issue of the IEO underscores important sources of uncertainty that can dramatically change the future course of events affecting trends in the composition, cost, and level of energy use. Economic growth and energy use are closely tied. Thus, assumptions about alternate paths of economic growth can have significant effects on expected energy requirements. The IEO99 reference case economic projection adopted for this outlook depicts a recovery in Asia’s economic growth beginning in 1999. With the exception of Indonesia, countries currently in recession in Asia are assumed to be back to baseline rates of growth within the next 3 to 5 years. Japan, too, is expected to resume positive rates of economic expansion by 2000. China and India are projected to sustain average growth rates in excess of 5 percent between now and 2020.

The events of the past year demonstrate that economic expansion can falter, with severe implications for energy market developments. The reference case economic growth assumptions may not be realized. Accordingly, alternative growth cases are included in the projections. More so than last year, the reference case projection depends on the continued efficacious application of economic policies designed to counter recession and foster structural reforms within national economies. In various countries, economic stress is leading to increased social tensions. Changing political forces could upset policy expectations in a variety of areas, causing reference case expectations not to be realized. Those who are more pessimistic or who have different views of the political and economic uncertainties in particular regions may have less optimistic expectations for economic growth in some of the regions covered in this review. The alternative growth cases presented here may provide useful perspective to assist their analyses.

IEO99 assumes that the reference case projections are the most likely outlook at this point in time. Aggressive policies to counter recession in Asia and to forestall spillover effects in Europe and the western hemisphere were instituted in 1998 by the International Monetary Fund (IMF), the World Bank, and banking institutions in Europe and the Americas. Signs of economic turnaround are now being reported for Korea, Thailand, Malaysia, and the Philippines.

Additional uncertainty arises from the commitments being made by developed countries under the Kyoto Protocol of the Framework Convention on Climate Change. If those commitments are realized, energy demand could actually be reduced over the next decade. Such a development could lower total world energy consumption in 2010 by more than 10 percent relative to the reference case projection presented here, equivalent to 60 quadrillion Btu or 30 million barrels of oil per day. The Protocol identifies stringent targets for reduced greenhouse gas emissions in developed countries; however, neither policies nor technologies to achieve the targets have been identified for implementation, nor have the countries ratified the agreement. Thus, the Kyoto Protocol is viewed in this report as a factor heightening uncertainties and the need for collateral analysis of the IEO99 reference case rather than one that per se alters it.

Another key source of uncertainty for the long-term evolution of energy markets relates to trends in energy intensities—i.e., the manner in which energy requirements evolve relative to growing income levels. History shows different trends in energy intensity for the developed and developing countries. In developed countries, energy requirements have grown slowly relative to increasing levels of economic activity. That historical trend is projected to continue in the reference case, with energy use rising at only about half the rate of economic expansion.

In the developing countries, energy and economic growth have tended to move in parallel. The process of economic development is energy intensive, and rising living standards enable broad access to electricity and motorized means of transportation. The accompanying widespread development of infrastructure causes growth in energy-intensive industries such as steel and cement. As economies continue to develop, however, the rate of energy use tends to fall relative to economic expansion. Consumer demands tend to evolve toward increased use of services that are not energy intensive.

The reference case projection presented here assumes a declining rate of energy intensity for developing countries over the projection period: by 2020, the relationship between energy and economic growth in developing countries is projected to be is similar to that in developed countries. This is a key assumption. Although per capita energy consumption is expected to rise over the next two decades, the projected levels in 2020 are still low relative to those currently prevailing in developed countries. If energy growth rates do not decline substantially relative to projected economic growth rates, the demand for energy in the developing world, where more than two-thirds of the world’s population resides, would be substantially higher.

Other key assumptions underlying the IEO99 reference case projection involve the continuing evolution of world oil markets. Current oil prices are near record low levels. Although some analysts now argue that low oil prices could persist for many years, the reference case assumes a price recovery path that begins in 1999 and returns to the levels projected in the IEO98 reference case by about 2007. Expectations for long-term recovery in demand growth and only a modest constriction on oil supply, especially in mature producing areas, encourage this view. In addition, IEO99 assumes no change in the influence of the Organization of Petroleum Exporting Countries (OPEC) on oil markets. It is also assumed that technology and government policies worldwide will support a large expansion in oil supply production capabilities.

The prospects for world energy consumption in IEO99 are somewhat lowered from last year’s report. Indeed, in this year’s projection, world energy consumption reaches 612 quadrillion British thermal units (Btu) in 2020, almost 30 quadrillion Btu less than last year’s forecast (Figure 9). The downward revision flows from events in two parts of the world: the Asian economic crisis—which began in the spring of 1997 and persisted throughout 1998, aided and abetted by the worsening economic situation in Japan—and the prolonged collapse of the Russian economy.

Sources: 1996: Energy Information Administration (EIA), International Energy Annual 1996, DOE/EIA-0219(96) (Washington, DC, February 1998). Projections: EIA, World Energy Projection System (1998 and 1999).

The persistent recession in Japan—the world’s second largest economy—has not shown signs of abating. In October 1998, the Japanese Parliament passed a $517 billion package to bail out the country’s national banking system and attempt to revive the economy [1]. The Japanese government has estimated that the economy will still shrink by 1.8 percent in the fiscal year ending March 31, 1999 [2]. While the Asian economic crisis does not appear to have dramatically affected China’s economy thus far, there are fears that the crisis could threaten its program of economic reform, which is designed to downsize state-owned enterprises in favor of more market oriented activities serving both domestic and export demands [3, p. 274].

Other Southeast Asian economies hit by the recession in mid-1997 began to fare better toward the end of 1998. Interest rates in the second half of 1998 fell sharply in South Korea and more modestly in Thailand, two countries that were hit particularly hard by the economic crisis. Both have adhered to the financial reforms prescribed by the IMF, and it appears that they might begin to see positive economic growth by the end of 1999 [4]. Some growth in exports is now evident, and an inflow of new foreign investment has begun. In contrast, the outlook for Indonesia is much less optimistic. By one estimate, the Indonesian economy was expected to contract by more than 13 percent in 1998, with no positive growth expected before 2000 [4, p. 4].

The deteriorating Russian economy—the largest economy in the former Soviet Union (FSU)—has led to some substantial downward revisions to the IEO99 projections. The lower expectations for energy growth are caused by the August 1998 devaluation of the Russian ruble, the defaults on public and private debt, the collapse of the Russian banking system, the worsening political situation for Boris Yeltsin and any potential successor, and expected changes in monetary policy that raise the possibility of hyperinflation. Less than one year ago, most forecasting sources were projecting positive growth in Russia’s GDP in 1998 and accelerating recovery in the years to follow. In October, PlanEcon revised its GDP forecast for Russia for 1999 from a positive 4.2 percent to a negative 5 percent, with no positive GDP growth expected before 2001 [5, 6].

For the FSU region as a whole, IEO99 projects that GDP will fall by an average of 2 percent per year between 1996 and 2000, with positive growth returning to the region between 2000 and 2005, averaging about 3 percent per year. As growth resumes, energy consumption in the FSU region is expected to recover. Between 1996 and 2020, energy consumption growth is expected to average 1.0 percent per year. The projection for 2020 is about 10 quadrillion Btu lower than in IEO98—equivalent to about 5 million barrels of oil per day.

Despite the downward revisions for expected energy demand worldwide, the IEO99 reference case projects that energy consumption in 2020 will increase by 236 quadrillion Btu—or about 65 percent—relative to the 1996 level. More than half the increment is expected in the developing countries, where strong economic growth in the long term is expected to increase the demand for energy over the projection period.

The predominant issue for the development of energy markets in the industrialized countries appears to be the potential impact of the Kyoto Protocol. As of March 15, 1999, 83 countries—including the United States—had signed the Kyoto Protocol, which calls for “Annex I” countries ³ to reduce or limit the growth of their carbon emissions between 2008 and 2012. The Protocol remains open for signature until March 15, 1999, but it will come into force only when 55 Parties to the Framework Convention, including Annex I countries that accounted for at least 55 percent of the total carbon dioxide emissions from that group in 1990, have “deposited their instruments of ratification, acceptance, approval or accession” [7]. As of March 15, 1999, only seven countries (Antigua and Barbuda, El Salvador, Fiji, Maldives, Panama, Trinidad and Tobago, and Tuvalu—all non-Annex I countries) had ratified the Protocol.

None of the Annex I countries had ratified the Kyoto Protocol by the time this report was prepared for publication. As a result, although the Protocol has the potential to change energy use dramatically in the industrialized world, no adjustments have been made to try to account for the impact of the Protocol. Nonetheless, the Protocol could retard the industrial world’s energy demand growth. The IEO99 reference case projections suggest that industrialized countries can expect to account for about 30 percent of the world’s increment in energy use between 1996 and 2010. Were emissions targets identified in the Protocol to be achieved by reducing fossil energy usage, energy consumption overall would be reduced by between 30 and 60 quadrillion Btu—equivalent to between 15 and 30 million barrels of oil per day. On the other hand, with potential fuel- switching opportunities, emissions trading, and other offsets (such as reforestation) allowed under the Protocol, a more modest reduction in fossil fuel use is more likely.

The strongest growth in energy consumption is expected to occur outside the industrialized world, which currently consumes about 40 percent more energy than is consumed by the developing world (Figure 10). By the end of the projection period, energy use in the developing countries (defined as developing Asia, Africa, the Middle East, and Central and South America) is expected to exceed that in the industrialized world. Such large increments in energy use in the developing world would have a dramatic effect on world energy markets. The IEO99 projections assume substantial financial investment in all phases of energy production, distribution, and transmission. If the assumed levels of investment are to be achieved, world government policies must continue to evolve to favor private-sector incentives for trade and development.

Figure 10. Projected Energy Consumption in the Industrialized and Developing Regions, 2000-2020

The IEO99 reference case, based on “business as usual” assumptions, projects that every energy source except nuclear power will grow over the 1996 to 2020 forecast period (Figure 11), although renewable energy sources are not expected to grow as fast over the next 24 years as they have in the past. Worldwide, oil remains the dominant source of energy throughout the projection horizon, as it has since 1970. Oil’s key role in the transportation sector—where it does not currently have any serious competition from other energy sources—helps to sustain its position among fuel sources. Oil use in the electric power sector is projected to decline in relative terms, but the fast-paced growth of personal transportation, especially in the developing world, will absorb any losses in the electricity sector.

Sources: History: Energy Information Administration (EIA), Office of Energy Markets and End Use, International Statistics Database, and International Energy Annual 1996, DOE/ EIA-0219(96) (Washington, DC, February 1998). Projections: EIA, World Energy Projection System (1999).

Natural gas is expected to be the fastest-growing primary energy source from 1996 to 2020. Worldwide consumption of natural gas increases by 3.3 percent per year (on a Btu basis) over the 24-year projection period, nearly twice as fast as oil (1.8 percent per year) and coal (1.7 percent per year). Gas is increasingly the fuel of choice for new electric power generation, primarily because combined-cycle gas turbine plants tend to be less expensive to build and are more efficient than other means of power generation. It is also a fuel of choice for environmental reasons. Local air pollution can be lessened by shifting from coal to natural-gas-fired generation. On a Btu basis, carbon emissions from natural gas combustion are less than half those for coal. Within the next decade, natural gas use is expected to exceed coal consumption, with the margin growing ever larger in subsequent years.

Coal use worldwide is projected to increase by 2.4 billion short tons, from 5.2 to 7.6 billion short tons, between 1996 and 2020. Strongest growth in demand is projected for the developing world, where coal use increases by 3.0 percent per year over the projection period (Figure 12). The worldwide increase in coal use is attributable mainly to increases in developing Asia—particularly, China and India. Indeed, IEO99 projects that China and India alone will account for more than 90 percent of the worldwide increment in coal consumption between 1996 and 2020. In the industrialized world, coal demand remains relatively flat through 2020, with average annual growth of 0.4 percent. Further, in Eastern Europe and the FSU (EE/FSU), coal consumption is projected to decline by 1.5 percent per year. There have been major declines in coal production and use in the EE/FSU since the social and political upheaval of the late 1980s and early 1990s. As the economies of the region recover, natural gas is expected to be used in place of those uses historically ascribed to coal.

Nuclear power is the only primary energy source projected to decline over the forecast period. After peaking at 2,390 billion kilowatthours worldwide in 2010, nuclear energy use is projected to decline to 2,068 billion kilowatthours in 2020. The worldwide decline is attributed to retirements of nuclear facilities in the industrialized world and in the FSU, where countries are operating older reactors and have other, more economical options for new generating capacity. In the developing world—especially developing Asia—increases in nuclear power generation still are planned in China, India, and South Korea. In addition, it is possible that ratification of the Kyoto Protocol could modify the outlook for nuclear power in the Annex I countries, where the operating lives of nuclear facilities could potentially be extended to constrain greenhouse gas emissions.

Hydroelectricity and other renewable resources maintain an 8-percent share of total energy consumption throughout the projection period. The growth of renewable resources is expected to be restrained somewhat by low fossil fuel prices, which discourage the development of renewable energy sources. As with nuclear power, ratification of the Kyoto Protocol could help renewable energy gain market share if the signatory countries used non-carbon-emitting energy sources to reduce their reliance on fossil fuels and consequently, reduce their greenhouse gas emissions. In Western Europe there is increasing activity in renewable installations, involving particularly wind-generated electricity. The German government announced in December 1998 that Germany’s renewable energy use expanded by 30 percent between 1996 and 1997, to 5 billion kilowatthours [8]; however, this still represents only about 1 percent of the nation’s total electricity consumption.

If energy consumption grows to levels projected in the IEO99 reference case, annual carbon emissions will reach 8.0 billion metric tons in 2010 and 9.8 billion metric tons in 2020 (Figure 13). Thus, world carbon emissions would exceed 1990 levels by almost 39 percent in 2010 and by 70 percent in 2020. Emissions are projected to rise by 2.2 billion metric tons between 1990 and 2010 and by another 1.8 billion metric tons between 2010 and 2020. Between 1990 and 2020, emissions from the combustion of coal are expected to account for 1.2 billion metric tons of the total increase in worldwide emissions, natural gas 1.5 billion metric tons, and oil 1.3 billion metric tons.

The Kyoto Protocol, if ratified and implemented, may well influence carbon emissions levels in the future. Only 2 of the 67 countries that had signed the Kyoto Protocol as of December 1998, Fiji and Tuvalu, have actually ratified it, and neither is required to reduce emissions levels under the treaty. As a result, IEO99 reference case projections have not been adjusted to account for changes that might occur under the Protocol. A discussion of how Kyoto-type scenarios might be achieved through fuel switching, reductions in fossil fuel consumption, and various emissions trading strategies is presented in the final chapter of this report, “Environmental Issues and World Energy Use.”

In the industrialized world, carbon emissions are projected to increase from 2.9 to 3.9 billion metric tons between 1990 and 2020, an increment of just over 1.0 billion metric tons (Figure 14). Emissions are expected to grow less than primary energy use mainly because of strong growth in the use of less carbon-intensive natural gas relative to coal, which is more carbon-intensive. About 52 percent of the total increase in emissions in the industrialized countries is attributed to an increase in the use of natural gas. Indeed, by the end of the projection period, emissions resulting from natural gas consumption exceed those from coal for these countries, reflecting the strong growth expected for natural gas relative to coal. Oil remains the dominant source of carbon emissions in the industrialized countries throughout the projection period, because of its role in transportation.

Figure 14. World Carbon Emissions in the Industrialized and Developing Regions by Fuel Type, 1990, 2010, and 2020

Sources: 1990: Energy Information Administration (EIA), International Energy Annual 1996, DOE/EIA-0219(96) (Washington, DC, February 1998). Projections: EIA, World Energy Projection System (1999).

Carbon emissions grow most quickly in the developing countries. By 2010 their emissions are expected to surpass those of the industrialized countries, whereas in 1990 developed countries’ emissions were nearly two times those of the countries of the developing world (Figure 15). By 2020, emissions in the developing countries are projected to exceed those of the industrialized countries by nearly 1 billion metric tons (in the context of a world total of 9.8 billion metric tons). The fast-paced growth of emissions in the developing world is attributed to high rates of economic and energy growth, as well as expectations for continued heavy dependence on fossil fuels (particularly coal, the most carbon-intensive of the fossil fuels) especially in developing Asia. Coal use is expected to contribute about 44 percent of the 3.2 billion metric tons of emissions added in the developing countries between 1990 and 2020.

In the EE/FSU region, carbon emissions remain below their 1990 levels throughout the projection period—a result of the economic and political upheaval in the region during the early 1990s. By 2010, emissions in the EE/FSU are expected to be about 27 percent lower than in 1990. Although some recovery in energy demand is projected by the end of the forecast period, emissions still are expected to reach only 1.0 billion metric tons by 2020, still about 21 percent lower than in 1990. The region is expected to be able to take advantage of its lower emissions levels if a worldwide carbon trading system is developed in the future. Industrialized Annex I countries may seek to buy emissions rights to meet their obligations under the Kyoto Protocol. Procedures for such a system have yet to be established among the participating countries, but negotiations in the matter continue.

Worldwide, carbon emissions per person increase from about 1.1 metric tons per person in 1990 to 1.2 metric tons per person in 2010 and 1.3 metric tons per person in 2020. Per capita emissions in the industrialized countries remain much higher than those of the rest of the world throughout the projection period, increasing from 3.2 to 3.8 metric tons per person between 1990 and 2020 in the IEO99 reference case. Even in 2020, with a doubling of per capita emissions, the developing world’s per capita emissions (0.8 metric tons per person) remain little more than one-fifth the level of the industrialized world. With four-fifths of the world’s projected population attributed to the developing countries in 2020, however, relatively small increments in per capita emissions would have a much greater impact on overall emissions than would larger increments in per capita emissions for the industrialized world.

Within the industrialized countries, the United States and Canada have the highest per capita emissions levels throughout the projection period—reaching 6.1 and 5.1 metric tons per person, respectively, in 2020 (Figure 16)—although the growth rate of per capita emissions in both countries is expected to remain fairly flat after 2000. In contrast, per capita emissions in developing countries such as China and India are projected to more than triple between 1990 and 2020, reflecting fast-paced industrialization based largely on fossil fuel consumption.

Figure 16. Carbon Emissions per Capita for Selected Regions and Countries, 1990-2020

Sources: History: Energy Information Administration (EIA), International Energy Annual 1996, DOE/EIA-0219(96) (Washington, DC, February 1998). Projections: EIA, World Energy Projection System (1999).

The pace of increase in carbon emissions per capita is even faster in the newly industrialized countries of Southeast Asia. In South Korea, for instance, per capita carbon emissions grew by 6.5 percent annually between 1970 and 1996, reaching 2.5 metric tons per person in 1996 (Figure 17). By 2020, per capita emissions in South Korea are projected to grow to 4.4 metric tons per person, matching the levels in Australasia and substantially exceeding those in Japan (2.9 metric tons per person). Oil remains by far the most widely used energy fuel in South Korea, maintaining a 63-percent share of total energy consumption over the next two decades. Were per capita emissions in China and India to grow over the projection period at the same rate as in South Korea over the past 26 years, total worldwide emissions could exceed current projections by 3.4 billion metric tons, assuming a continuation of current fuel use patterns.

IEO99 includes a high economic growth case and a low economic growth case in addition to the reference case. The reference case projections were derived by establishing a set of regional assumptions about economic growth paths and energy elasticity (the relationship between changes in energy consumption and changes in GDP). The two alternative growth cases, based on alternative ideas about the possible paths of economic growth, were formulated to provide users with a way to quantify the range of uncertainty associated with the reference case.

For the high and low economic growth cases, different assumptions were made about the range of possible economic growth rates in developing and industrialized nations, reflecting the greater uncertainty inherent in attempts to forecast economic growth in developing economies. The same pattern of change in energy intensity relative to GDP (discussed below) was assumed for the high and low growth cases as in the reference case. For industrialized countries, increments of +1.0 and -1.0 percentage points, respectively, were added to the reference case growth rates to generate high and low growth cases. For nonindustrialized countries and regions (apart from China and the EE/FSU), increments of +1.5 and -1.5 percentage points were used to generate the high and low growth cases.

China and the EE/FSU are special cases with regard to prospects for future economic growth. China has experienced high economic growth in the past several years, and the EE/FSU region has suffered a severe economic downturn. In both regions, there is opportunity for substantial change in growth: China has the potential for a larger decline in growth rate given its currently high rate, and there are prospects for a substantial increase in the rate of growth for the EE/FSU nations should their current political and institutional problems be moderated enough for the recovery of a considerable industrial base. Reflecting these uncertainties, -3.0 percentage points were added to China’s growth rate for the low economic growth case and +1.5 for the high case; and +3.0 percentage points were added to the EE/FSU growth rate for the high economic growth case and -1.5 for the low case.

In the IEO99 reference case, total world energy consumption is expected to reach 612 quadrillion Btu in 2020, with the industrialized countries consuming 263 quadrillion Btu and the rest of the world 349 quadrillion Btu. Under the assumptions of the high economic growth case, total world energy consumption would be 747 quadrillion Btu in 2020, 136 quadrillion Btu higher than in the reference case (Figure 18). In the low economic growth case, worldwide energy consumption in 2020 would be 497 quadrillion Btu, 115 quadrillion Btu less than in the reference case. Clearly, there is a substantial range between the low and high economic growth cases. The range between the cases for total world energy consumption—251 quadrillion Btu—is more than 40 percent of the total reference case consumption projected for 2020.

Another way to look at uncertainty in long-term energy demand developments is in terms of the way energy demand evolves relative to GDP over time. Economic growth and energy demand are linked, but the strength of that link varies among regions and stages of economic development. In industrialized countries, history shows the link to be relatively weak. That is, energy demand growth lags behind economic growth. For every percent increase in economic activity, energy demand increases only about half a percent. In developing countries, demand and economic growth have tended to be more closely correlated, with energy demand growth tending to track the rate of economic expansion.

The historical behavior of energy intensity in the FSU is problematic. The EE/FSU economies have always maintained higher levels of energy intensity than either the industrialized or developing nations (Figure 19). In the FSU, however, energy consumption grew more slowly than GDP until 1990, when the collapse of the Soviet Union created a situation in which both income and energy use were declining, but GDP fell more rapidly. As a result, energy intensity began to rise.

The stage of economic development and the standard of living of individuals in each region strongly influence the link between economic growth and energy demand. Advanced economies with high living standards tend to have relatively high energy use per capita, but they also tend to be economies where per capita energy use is relatively stable or changes occur very slowly. Here increases in energy use tend to correlate with employment and population growth. There is a high penetration rate of modern appliances and motorized transportation equipment in the industrialized countries. As a result, increments to personal income tend to result in spending on goods and services that are not energy intensive. To the extent that spending is directed toward energy-consuming goods, it involves more often than not purchases of new equipment to replace old capital stock. The new stock is often more efficient than the equipment it replaces, resulting in a weaker link between income and energy demand. In developing countries, standards of living, while rising, tend to be low relative to those in more advanced economies. As a result, many energy-using devices are being widely used for the first time, causing energy use to track more closely with rising income levels.

Indeed, changing growth patterns of energy intensity could have dramatic impacts on energy consumption in the projection period, particularly among the developing countries. For instance, if energy intensities are assumed to decrease (improve) in the developing countries according to a percentage that represents the single greatest annual improvement observed between 1990 and 1996, energy consumption in the developing world would be 174 quadrillion Btu in 2020—about 100 quadrillion Btu less than the reference case projection of 279 quadrillion Btu. On the other hand, if energy intensities in the developing world are assumed to increase (worsen) by the highest annual rate of the 6-year period, energy consumption in the developing countries climbs to 926 quadrillion Btu in 2020—more than three times the reference case projection.

Another way to illustrate the uncertainty associated with the projections presented in the IEO99 is to compare them with those derived by other forecasters. Four organizations provide forecasts comparable to those in IEO99. The International Energy Agency (IEA) provides “business as usual” projections to the year 2020 in its World Energy Outlook 1998. Standard & Poor’s DRI (DRI) also provides energy consumption forecasts to 2020 in its World Energy Service: World Outlook 1998. Petroleum Economics, Ltd. (PEL) and Petroleum Industry Research Associates (PIRA) publish energy forecasts for the world, but only to the years 2015 and 2010, respectively.

All the forecasts for total world energy consumption are similar (Table 3). Both IEA and DRI project growth of 2.0 percent per year between 1995 and 2020, and IEO99 projects 2.1-percent annual growth. The PEL and PIRA growth rate projections are similar, but their projection horizons are not the same as those for the other forecasts. The PEL projections are for 1995 to 2015, and the PIRA projections are for 1995 to 2010. For those periods, the IEO99 reference case projects 2.1-percent annual growth from 1995 to 2015 (compared with PEL’s projection of 2.0 percent) and 2.2-percent annual growth from 1995 to 2010 (the same as the PIRA projection).

In the projections that extend through 2020, almost all the projected growth rates for the industrialized region fall within the range defined by the IEO99 low and high economic growth cases. One notable exception is the IEA projection for North America. IEA expects energy use in North America to increase by only 0.7 percent per year between 1995 and 2020, whereas the IEO99 reference case projects growth of 1.3 percent per year. The IEA suggests that the differences between its projections for North America and the IEO99 projections may be attributed, in part, to its expectations of higher oil and natural gas prices [3, pp. 222-223].

Comparing projections from the PIRA series with those from IEO99 over the 1995-2010 period, North America is the only industrialized region for which there are substantial differences between IEO99 and an alternative series. PIRA expects energy use in North America to grow by 1.2 percent per year between 1995 and 2010, whereas the IEO99 reference case projects 1.5-percent annual growth. In the IEO99 low economic growth case, North America’s energy consumption increased by 1.2 percent per year, exactly the same as the rate in the PIRA series. Comparing IEO99 and PEL over the 1995-2015 period, the PEL growth rates for all industrialized regions fall within the range defined by the IEO99 low and high cases.

The FSU region provides severe challenges for those developing long-term projections. Only one other forecast, by PEL, provided projections for the FSU separated from the Eastern European countries. Over the 1995-2015 period, IEO99 frames a wide range for projected growth in energy use in the region, ranging from 0.1 percent per year in the low economic growth case to 2.1 percent per year in the high economic growth case (the 1995-2020 time frame presented in Table 3 provides a range between 0.2 percent per year in the low growth case and 2.4 percent per year in the high growth case). Nevertheless, PEL projects even worse performance than in the IEO99 low growth case, with no expected growth between 1995 and 2015. For the entire EE/FSU region, all the forecasts have energy consumption growth rates that fall into the range defined by the IEO99 low and high economic growth cases, including comparisons of PEL and IEO99 cases over the 1995-2015 period and of PIRA and IEO99 over the 1995-2010 period.

As might be expected given the continuing regional economic recession, Asia is also a region where the expectations for energy demand vary considerably. In the IEO99 reference case and in the IEA ‘business as usual’ case, energy demand in developing Asia is expected to increase by 3.7 percent per year between 1995 and 2020. DRI projects a lower growth rate of 3.2 percent per year. PIRA projects a growth in energy use of 3.9 percent per year between 1995 and 2010 (the IEO99 reference case projects the same growth rate over the same time period), and PEL projects a growth rate of 4.1 percent per year between 1995 and 2015 (IEO99 is somewhat lower at 3.8 percent per year over the same period). Other Asia includes countries hardest hit by the economic troubles that began in 1997 and continued throughout 1998, such as Thailand, Indonesia, Malaysia, and South Korea. It is clear that there is still much debate among analysts about the time frame needed for these countries to regain momentum for economic expansion and increased demand for energy.

Within Asia, there is more variation in expectations for “other Asia” than for China. IEO99 and DRI tend to be more conservative than the other forecasts. Between 1995 and 2020, IEO99 and DRI project energy demand growth in other Asia of 3.3 percent and 3.0 percent per year, respectively. IEA projects 4.0-percent annual growth in other Asian energy use. Between 1995 and 2015, PEL expects energy use in other Asia to increase by 4.1 percent per year, compared with 3.4 percent per year for the same period in the IEO99 reference case. Between 1995 and 2010, PIRA projects robust average annual growth of 4.4 percent, compared with 3.6 percent in the IEO99 reference case.

Of the remaining developing regions, the greatest variation in expected growth is seen for Central and South America. IEO99 is more optimistic about growth in energy use in the region than are any of the other forecast series, projecting growth of 4.3 percent per year over the 1995-2020 period, compared with 3.1 percent per year (DRI) and 2.9 percent per year (IEA). Over the 1995-2010 period, IEO99 is substantially higher than PIRA in terms of energy demand growth (4.5 and 3.6 percent per year, respectively); and over the 1995-2015 period, IEO99 projects a substantially higher growth rate (4.4 percent per year) than does PEL (3.6 percent per year).

A key reason for the differences among the various forecasts is that they are based on different expectations about future economic growth rates (Table 4). Expectations for economic growth are substantially alike among estimates for the industrialized regions from most of the forecasts. The IEA economic growth rates for North America, Western Europe, and the Pacific are lower than the corresponding IEO99 growth rate projections. The PIRA and PEL forecasts have substantially the same economic growth rates as IEO99 for the comparable time frames.

Projected GDP growth rates over the 1995-2020 period vary somewhat for the EE/FSU region, as might be expected given the economic straits in which the FSU, and specifically Russia, finds itself. While IEO99 assumes annual GDP growth of 2.9 percent for the region, IEA projects a more optimistic 3.3 percent per year. The projections in last year’s IEO were more optimistic at 3.7 percent. Over the 1995-2010 period, PEL and IEO99 are in relative agreement on the expectations for economic growth in the FSU. PEL projects GDP growth averaging 1.4 percent per year and IEO99 1.3 percent per year. The PEL projection for Eastern European countries is more pessimistic than IEO99. PEL projects growth of 3.2 percent per year between 1995 and 2010. Over that same time period, IEO99 projects Eastern European GDP growth at 4.5 percent.

Projections vary not only with respect to the levels of energy demand and economic growth but also with respect to the composition of energy input use (Table 5). IEO99 expects continued strong growth in world natural gas consumption, growing by 3.3 percent annually between 1995 and 2020. This growth rate, which is the same as the IEO98 projection, is the most optimistic of all the forecasts. Growth rates for natural gas use among the alternative projections range from 2.3 percent per year (PEL) to 2.9 percent per year (PIRA).

In IEO99, projections for all energy sources except natural gas and nuclear power were revised downward from last year’s report—a result of increased pessimism for recovery of the economies of the FSU and a downward revision based on the impact of the Southeast Asian economic crisis. The decline projected in IEO98 for world nuclear consumption was revised upward slightly, so that nuclear energy is now projected to decline by 0.3 percent per year between 1995 and 2020 rather than 0.4 percent per year.

Given the relatively high growth rate of expected gas use, IEO99 tends to have lower growth rates than the alternative forecasters for the remaining fossil fuels. In IEO99, oil use grows by 1.8 percent per year worldwide between 1995 and 2020, whereas the range of average annual growth rates among the other forecasters runs from 1.9 percent (IEA) to 2.0 percent (DRI). Coal use in IEO99 grows by 1.7 percent per year, but the DRI and IEA forecasts project increases of 2.1 percent per year. For both oil and coal, the alternative forecasts fall within the range of projections defined by the IEO99 low and high economic growth cases.

At first glance, it appears that the IEO99 projections of growth in the use of non-fossil fuels (i.e., nuclear power and hydroelectricity and other renewables) also are lower than those in the other forecasts. However, DRI produces a forecast only for combined nuclear and hydroelectricity—as “primary electricity.” When the IEO99 projections for the two non-fossil energy sources are combined, they fall in the middle of the range of growth rates expected between 1995 and 2020 by the other forecasters, from 0.7 percent per year (DRI) to 1.2 percent per year (IEO99).