International Energy Agency, 9 rue de la Fédération, 75739 Paris Cedex 15, France

Abstract

This paper provides an overview of the IEA's 1998 edition of the World Energy Outlook. It discusses the likely developments in global energy demand for the period to the year 2020. Special attention is paid to the power generation sector. Prospects for electricity generation are discussed in detail for 10 world regions, followed by a description of the model that was used to produce these projections. Finally, the paper discusses the main uncertainties surrounding the projections presented here.

Article Outline

1. Introduction

2. Economic growth and price assumptions

3. Energy demand projections

4. CO₂ emissions

5. Outlook for the power generation sector

6. Regional highlights

7. Generating capacity

8. Fuel use in power stations

9. The power generation model

10. Uncertainties

11. Comparison with other forecasts

11.1. World

11.2. OECD North America

11.3. OECD Europe

11.4. OECD Pacific

References

1. Introduction

The 1998 edition of the IEA's World Energy Outlook [1] aims to identify and discuss the main issues and uncertainties affecting world energy demand and supply over the period to the year 2020. It does so in the framework of a business as usual (BAU) projection which assumes energy policies existing before the Kyoto Conference of December 1997 remain in place and that no new major policies are adopted to reduce energy-related greenhouse gases.

The outlook provides detailed demand and supply balances and CO₂ emissions for 10 world regions.¹ It projects that world energy demand will grow by 65% and CO₂ emissions by 70% between 1995 and 2020 unless new policies are put in place. It assumes a rate of world economic growth of 3.1% p.a. (1990 US dollars and purchasing power parity²), close to the actual rate since 1971. Two-thirds of the increase in energy demand over the period 1995–2020 arises in China and the other developing countries.

2. Economic growth and price assumptions

The projections of energy demand are based on a set of assumptions on future economic growth and international fossil fuel prices.

Economic growth is arguably the most important driver of energy demand. The BAU assumption broadly continues the past world rate of economic growth. All regions are expected to experience slower growth in the future, except for the Transition Economies which are assumed to recover rapidly from the economic turmoil of the 1990s. As the shares of the rapidly growing developing countries are rising, the world average growth rate remains close to its past level (Fig. 1).

World primary fuel price assumptions for the BAU projection are plotted in Fig. 2. In the period to 2010 the price of crude oil is assumed to remain flat at the average 1991–95 levels. It is increased between 2010 and 2015 to reflect the expected transition from conventional to unconventional oil as the source of marginal supply.

The prices for natural gas in Europe and LNG in Japan are increased in the same proportion to reflect the close competition with oil products. The world price of coal is also increased to take account of the corresponding increase in transport costs.

The US natural gas wellhead price is increased over the period 2005–2015 to reflect a possible tightening of the North American gas market and increased use of unconventional gas.

3. Energy demand projections

World energy demand is projected to increase at an average annual rate of 2% over the outlook period. This is slightly lower than the growth rate of 2.2% per annum experienced from 1971 to 1995. In absolute terms, the annual global demand for energy increases from 8341 Mtoe in 1995 to 13,749 Mtoe in 2020. These figures do not include biomass energy consumption in developing countries. Projections of worldwide biomass are available in the 1998 edition of the World Energy Outlook, but this paper focuses on commercial energy trends only.³

A key message of this outlook is that fossil fuels will continue to dominate the energy mix. Ninety-five percent of the additional energy demand between 1995 and 2020 will be met by fossil fuels. Projections of world energy demand by fuel are illustrated in Fig. 3.

Oil continues to dominate world energy consumption, although its share declines from 10% in 1995 to 8% by the end of the outlook period. Most of the increase in oil demand will stem from additional demand for transportation services.

Demand for gas increases faster than that for oil or solid fuels, particularly in the OECD regions. Where pipelines exist, or can be put in place, natural gas is the preferred fuel for many applications, especially for new power stations. Gas consumption nearly doubles over the outlook period, rising from 1810 Mtoe in 1995 to 3468 Mtoe by 2020. The share of solid fuels in the primary energy mix will remain largely unchanged over the projection period. Nearly three-quarters of additional demand for solid fuels will be in the power generation sector. Growth is faster in the developing regions, notably in China and South Asia which are expected to develop further their indigenous supplies of coal.

World nuclear power stabilises. In the OECD, some new nuclear plants are expected to be built during the outlook period. At the same time, several plants will reach the end of their operational life leading to an overall decline of nuclear power in the region. Outside the OECD, growth could be higher, with nuclear power increasing in Asia and in the Transition Economies.

Hydro power is expected to grow at an average rate of 2% per annum. Within the OECD, most of the sites have already been exploited and therefore growth is expected to be limited, at 0.7% per annum. Developing regions endowed with hydro resources will increase the use of this energy form to generate more electricity. China and East Asia are the regions where growth in hydropower is expected to be the highest.

The use of other renewable energy sources increases steadily, but remains at low levels. Although renewable energy is projected to have the highest growth rate over the outlook period, it does so from a very low basis and therefore its share by 2020 will still be less than 1% of the primary energy mix. The main reason why this share remains low is that the cost of developing renewable energy, under business as usual assumptions, remains high over the projection period. The power generation sector will be the primary consumer of this form of energy. Among the different forms of renewable energy (wind, geothermal, solar, tidal), generation from wind power is expected to make the largest contribution.

4. CO₂ emissions

Over the period 1971–95, CO₂ emissions grew at an average rate of 1.7% per annum. The outlook projects a faster growth rate of CO₂ emissions for the period to 2020, at 2.2% per annum. Contributing factors are the stabilisation of nuclear power generation and continued rapid growth in coal use in China and other Asian countries. By 2020, the developing countries could account for half of global CO₂ emissions. Table 1 provides data and projections of CO₂ emissions for the four major regional groupings presented in the outlook.

5. Outlook for the power generation sector

Electricity is the fastest growing component of energy demand worldwide. Over the period 1971–95 growth averaged 3.8% per annum, which is higher than the 3.2% average annual growth rate of global economy in the same period.

This paper presents a Business-As-Usual projection of electricity generation at a world level. It shows that world electricity demand grows fast, at 3% per annum over the period to 2020. The main assumption that underlies this projection is that at the same time the world economy will grow at 3.1% per annum (measured in purchasing power parities). Natural gas is expected to double its share in the electricity generation mix, from 15% currently to 30% by 2020.

In 1995, world electricity generation amounted to 13,204 TWh. With demand for electrical services growing rapidly, the annual level of generation could more than double by 2020. Generation grows strongly in all regions and is dominated by the OECD, whose world share, nevertheless, is projected to fall from 60% in 1995 to 47% in 2020. The generation share of China rises from 8% to 14%, for the Transition Economies it remains at 12% and for the Rest of the World, it rises from 19% to 27% over the period.

Table 2 shows the IEA's projections of world electricity generation as well as projections of generation by energy source.

Solid fuel, mainly coal, retains a strong position in power generation. It is the favoured fuel where gas is unavailable or expensive (as in those developing countries that have coal available, like China and India), or in locations close to low-cost coal production (parts of North America, Australia and South Africa). The main threat to coal use in power generation comes from future policies to reduce emissions of CO₂.

Oil use in power generation grows to 2020, but less quickly than total generation. Because of the relative ease and low cost of oil storage, it is an ideal generating fuel for remote locations where other fuels are difficult or costly to obtain, for standby or peaking plants and for use where seasonal variations in price make other fuels (especially gas) uncompetitive at certain times.

Electricity generation from gas increases rapidly, at an annual average rate of 6% during the projection period. By 2020, gas becomes the second largest source of electricity, at a world level.

World nuclear power increases slightly to 2010 and then starts declining. In the OECD, some new nuclear plants are expected to be built during the outlook period. At the same time, several plants will reach the end of their operational life leading to an overall decline of nuclear power in the region. Outside the OECD, growth could be higher, with nuclear power increasing in Asia and in the Transition Economies.

Hydro power is expected to grow at an average rate of 2% per annum. Within the OECD, most of the sites have already been exploited and therefore growth is expected to be limited, at 0.7% per annum. Developing regions endowed with hydro resources will increase the use of this energy form to generate more electricity. China and East Asia are the regions where growth in hydropower is expected to be the highest.

The use of other renewable energy sources increases steadily, but remains at low levels. Although renewable energy is projected to have a very high growth rate over the outlook period, it does so from a very low basis and therefore its share by 2020 will still be less than 1% of the electricity generation mix. Among the different forms of renewable energy (wind, geothermal, solar, tidal), generation from wind power is expected to make the largest contribution, particularly in the OECD.

6. Regional highlights

In OECD Europe, demand for electricity is expected to increase by 2.1% per annum over the period 1995–2020. By 2020, the electricity generation mix in OECD Europe is projected to be quite different from that of today. Most of the incremental demand for electricity is expected to be met by natural gas, and its share in generation increases rapidly, from 10% at present, to 45% in 2020. Coal and nuclear power, which today are Europe's most important sources of electricity supplying more than 60%, will decline to 34% in 2020. Significant increases are expected in renewable energy (other than hydro) use for electricity generation, notably in wind power, which could grow from 4 TWh in 1995 to 66 TWh in 2020. The share of non-hydro renewable energy for electricity generation will still remain low, at about 1.8% of total electricity generation in 2020.

Electricity generation in North America increases at an average rate of 1.8% per annum. By 2020, annual generation could represent 6362 TWh, a 55% increase above current levels. Coal and gas are expected to be the key fuels in the projected electricity mix. Gas-fired generation appears to be the most economic option for a new plant, particularly in the first half of the outlook. In the second half, higher gas prices, associated with higher gas production costs, could switch the economics of power generation in favour of coal. Consequently, although the share of natural gas generation increases from 13% in 1995 to 24% of the total in 2020, coal will continue to supply the bulk of electricity in the region. Nuclear electricity decreases, as no new nuclear plants are built and nuclear plant retirements accelerate in the second half of the outlook period. Hydropower shows a small increase, and other renewables increase slightly, as their costs remain high.

By the end of the projection period, annual power plant output in the OECD Pacific region, could be 57% above its 1995 level in line with the 1.8% per annum growth projected for electricity demand. Electricity output from gas-fired power stations could increase its share in the output mix by 10 percentage points over the outlook period, rising from 18% currently to 28% by 2020. Most gas in the region is traded in the form of LNG, which is an expensive form of energy (Japanese LNG import prices are 30–40% higher than the border price of gas to Europe via pipeline). LNG competes with coal in the region, but by 2020, gas-fired generation overtakes that of coal. Nuclear power, concentrated in Japan, shows significant increase and becomes the largest single source of electricity by 2020; its share of total generation increases from 24% in 1995 to 28% in 2020.

In the Transition Economies, economic growth is assumed to resume and consequently, demand for electricity is expected to grow at an average annual rate of 2.9%. Significant expansion of electricity generation from gas should be expected as gas supplies from Russia and the Caspian region become increasingly available. Coal will remain an important fuel for electricity generation, although it is projected to lose some of its share to gas. Nuclear power is assumed to increase to 2010 and to start declining afterwards, as several nuclear power stations are decommissioned in that period and capacity additions are not expected to keep pace with retirements.

In China, electricity generation grows at 5.4% per annum to reach 3857 TWh in 2020. The power sector will continue to rely on coal, although coal's share in the electricity generation mix is projected to decline to two-thirds, as some incremental demand for electricity is covered by nuclear and hydro plants. Electricity generation from coal is projected to increase at 5% per annum. Nuclear generation increases from 13 to 127 TWh by 2020 and hydro from 191 to 726 TWh.

Much of the power sector expansion in East Asia is expected to rely on gas and coal. Total electricity generation in this region is projected to increase by 4.9% per annum in the period to 2020. Increased use of gas would require expansion of pipelines or LNG infrastructure, but these may be delayed because of lack of funding, given the current crisis in the region. Coal plants may be preferred if gas is unavailable or if indigenous supplies are reserved for export outside the region.

Electricity generation in South Asia is projected to increase from 485 TWh in 1995 to 1657 TWh in 2020. The region is expected to remain dependent on coal-fired generation which is projected to increase at 5.2% per annum and to reach 1024 TWh by 2020. Hydro-electricity is projected to double over the period 1995–2020, but as resources are depleted, growth in hydro is slowing down and its share in total generation is decreasing. Some growth is expected in natural gas fired generation, particularly in the second half of the outlook period, when current plans to expand gas fields, gas pipelines and LNG terminals could come on line. Electricity generated in nuclear power plants is assumed to rise to 19 TWh by 2020, up from 8 TWh in 1995. Almost all of the increase will come from new nuclear reactors in India. Alternative energy resources are also being explored in the region, particularly in India. Generation from renewable energy could increase from close to 0 level in 1995 to about 10 TWh in 2020.

Electricity generation and capacity in Latin America are projected to grow at 4% a year. Over the outlook period, the region's electricity-generation mix could change significantly and become more dependent on fossil fuels, notably on gas. Exploitation of the region's abundant natural gas resources will give a boost to natural-gas fired generation which is projected to increase 8-fold by 2020. Hydropower, currently the most important source of electricity generation, is assumed to double between 1995 and 2020; the rate of growth will slow down as the best sites are developed.

In the Middle East, annual electricity generation increases at an average annual rate of 3.8% over the outlook period. The electricity mix, currently dominated by oil and gas, is expected to become more dependent on gas as countries in the regions seek to free oil for export. By 2020, the share of gas in electricity output is expected to rise to 60%, up from 44% in 1995. The share of oil, on the other hand, is projected to decline, from 46% in 1995 to 28% in 2020.

Electricity output in Africa is projected to grow at 3.4% per year over the period to 2020. Coal is expected to lose market share, but will still generate about 43% of all electricity in 2020. Use of natural gas in the power sector increases rapidly. Most of the growth in gas use will occur in the countries of North Africa.

7. Generating capacity

Between 1995 and 2020, world power generating capacity is projected to increase from 3079 to 5915 GW. New capacity requirement over the period is 3503 GW, including some 667 GW of existing plant expected to be retired over this period. Significant retirements are expected to take place in the Transition Economies, in OECD Europe and in OECD North America, where some of the oldest plants are located. About a third of the new capacity is projected to be built in the OECD region and about one-half in China and the developing countries.

Table 3 shows that most new generating plants use gas, mainly in gas-fired combined-cycle turbines (CCGT). By 2020, one-third of world electricity generating capacity will be gas-fired. CCGT plants have low construction cost (currently in the range of US$ 450–650 per kW), are available in a range of small to medium sizes, have short construction times (2–3 years), and are straightforward to build and operate. They also have high efficiency and low pollutant emissions. Provided gas is available at a competitive price, they are expected to be the generating plant of choice over the projection period.

Over the outlook period, significant investment has to be made in new power generating schemes. The total capital expenditure needed for new capacity (excluding new transmission lines) for the 25-year period of the outlook is estimated to be in the order of US$ 3.28 trillion, i.e. an average capital cost of US$ 937 per kW of capacity. This average capital cost reflects a range of new plant capital costs. Although most of new power plants will be gas-fired CCGTs with low capital costs as mentioned above, a number of new plants will be using coal, nuclear or hydro power to generate electricity. These plants have much higher capital costs than CCGTs (for example, a coal-fired power plant costs US$ 750 per kW in China, and as much as US$ 2100 per kW in Japan), hence the high average value of capital cost for a new plant.

8. Fuel use in power stations

Fuel inputs to power stations increase by 2.3% per annum over the outlook period. This growth rate is lower than the projected growth rate for electricity output, because of significant improvements in power plant efficiency. Much of the increase in efficiency is attributed to the projected surge of natural gas use in CCGT plants which offer the highest conversion efficiency amongst all commercially available technologies. Indeed, output from gas-fired power stations increases by 6% per annum, while at the same time fuel input increases at an annual rate of 3.9% per annum. Inputs to coal-fired power stations increase by 2.5% per annum (compared with 2.9% annual increase in output). Oil consumption increases by 1.2%, against 1.6% increase in oil-fired generation. Table 4 shows projections of fossil fuel consumption in power stations. The table also shows for each fuel the share of power sector in total demand.

9. The power generation model

The power generation model provides a simple, quantitative framework within which the many issues that arise in the sector may be analysed and quantified. The model uses the electricity demand projections and assumptions for fossil fuel prices, capital and operating costs and efficiencies of new plants to calculate the amount of new capacity, electricity generated and the fuel consumed by type of plant.

The calculation simulates, in a simplified manner, the way these outputs are determined in practice. For market economies this means a lifetime, least-cost calculation for choice of new generating plants and a short-term, least-cost calculation for dispatching existing generating plants. Where plant or fuel price data do not exist, or where the choice of new plant or plant dispatching decisions are not made on a cost-minimising basis, more judgemental methods must be used, based on analysis of the data that do exist, on a review of the literature and on consultations with experts in the regions concerned. Because investments in nuclear and renewable plants have costs that are highly site and country specific and are frequently determined on a semi-political basis, they are determined by assumption.

The demand for electricity is combined with an assumed load curve to calculate peak load. The need for new generating capacity is calculated by adding a minimum reserve plant margin to peak load and comparing that with the capacity of existing plants less plant retirements using assumed plant lives. An allowance is needed for assumed plant availability.

If a new plant is needed, the choice is made on the basis of levelised cost. This is a technique widely used in the power sector by other modellers. The levelised generating cost (expressed as money value per kWh) combines capital, operating and fuel costs over the whole operating life of a plant using a given discount rate and plant utilisation rate.

Once the existing set of plants has been determined, fossil fuel prices are used to load plants in ascending order of fuel and operating cost, allowing for assumed plant availability. Once the generation of each type of plant has been determined, the fuel requirements are calculated using plant efficiencies.

One of the limitations of the power generation model is that it considers each region as a single utility with a single price for each generation fuel. This is seldom the case for a single country, and is certainly not the case for any of the world regions considered in the World Energy Outlook. For example, in OECD Europe, there are many variations in fuel prices across the 21 countries of the region. Coal prices are much higher in Germany and gas prices are generally lower in The Netherlands and in Italy than in other European countries. Adjustment for such regional differences in fuel prices are made in the projection both in the merit order calculation and in the choice of new plant, as well as in calculating the average prices for generating fuels for the region.

The future values of capital and operating costs of different types of generating plants and their efficiencies are not known with precision and may vary from one country to another within a region. For this reason, formal cost-minimising calculations must be treated with caution.

10. Uncertainties

Over the next 20–25 years demand for electricity is expected to grow very fast, particularly in developing regions. There are many uncertainties surrounding the projections of electricity demand and supply presented here. A major one is the pace of restructuring and deregulation of the electricity industry that is taking place in many countries. Expectations are that these developments will lead to greater efficiency and to electricity prices reflecting more closely the full costs of supply.

In developing countries, the expansion of existing power generation systems will require substantial capital expenditures. In the absence of new policies, many of these countries will have difficulty generating sufficient investment funds to carry out the necessary expansion. Their ability to finance the required large additions to electricity generating capacity is one of the major uncertainties of the outlook. It may be necessary for these countries to attract foreign funds which, in turn, may require the creation of a more deregulated environment.

Another major uncertainty concerns future policies to reduce greenhouse gases, following the Kyoto Protocol and countries' commitments. The packages of policies that individual countries will adopt to meet their commitments have not yet been determined. These policies, including the future role of nuclear power, may well affect the future electricity trends presented in this paper.

11. Comparison with other forecasts

Projections of energy demand are also available by other organisations. A comparison between these projections and the World Energy Outlook is provided below for the world and for the three OECD regions.

11.1. World

Table 5 provides a comparison of world energy demand projections between the US Department of Energy International Energy Outlook 1998 (US-DOE) [2] and the 1998 edition of the IEA's World Energy Outlook (WEO 98).

The two sets of projections assume the same world GDP growth rate, 3.1% per annum over the period 1995–2020. However, the US-DOE assumes lower fuel prices than the IEA and this explains, to a large extent, their projected higher growth rates for energy demand. In the US-DOE projection, oil prices rise to about US$ 19 per barrel (1990 values), while in the World Energy Outlook the price of a barrel of oil reaches US$ 25 (1990 values).

11.2. OECD North America

The World Energy Outlook projections for North America are compared here with the corresponding US-DOE projections. Both organisations assume the same GDP growth rate for the period 1995–2020.

Over the projection period, the US-DOE reference case for OECD North America foresees a higher pace of energy demand growth than does the World Energy Outlook. As shown in Table 6, US-DOE projects an average growth of primary energy demand of 1.2% per annum against 0.8% for the World Energy Outlook. In the IEA projection, the assumed higher energy price environment leads to a slowdown in energy demand, in particular that of oil and gas. The main difference between US-DOE and WEO 98 assumptions lies in the evolution of gas prices. Because of assumed rising North American gas production costs, the projections in the World Energy Outlook assume significantly higher natural gas prices than is assumed in the US-DOE's projection.

The differences with OECD North America oil demand projections are mainly due to expected developments in the transportation sector. The higher oil prices in the IEA model lead to a lower growth rate for the sector.

For power generation, the assumed increases in the gas price leads to less use of gas and consequently more coal in the IEA's projection. This is reflected in the higher growth rate of solid fuels in total primary energy demand.

11.3. OECD Europe

Apart from the IEA, two other organisations produce European energy demand projections: the European Union (EU) and the United States Department of Energy (US-DOE). The last set of energy projections produced by the EU was in 1996 and the Conventional Wisdom (CW) scenario [3] is used here for comparison purposes. The US-DOE reference case [2], is the other projection considered here. Note, however, that the geographic coverage of the region is not exactly the same in the three sets of projections.

The important point to note is that the EU appears to be assuming a much greater decline in the ratio between energy and GDP. In the EU projections, the energy ratio declines at an annual average rate of 1.5% compared to 0.9% in the WEO projection after 1994. Energy prices, assumptions about new technologies and changes in consumer behaviour appear to play a significant role in the EU's projections.

Table 7 compares the projected growth in the European energy demand during the period 1995–2020 for the World Energy Outlook, the European Union Conventional Wisdom scenario and the US-DOE International Energy Outlook.

The IEA and US-DOE project energy demand to grow at a similar rate of 1.1–1.2%, considerably higher than the EU's 0.7%. Both the US-DOE and the EU have lower projections than the IEA for oil and nuclear power. An interesting feature of the US-DOE projection is the high projected growth rate in gas demand of 3.8%. This is, in part, because gas replaces oil in heat production.

11.4. OECD Pacific

There are three other international organisations that produce long-term energy projections for the OECD Pacific region; the Asia Pacific Energy Research Centre (APERC), the European Union (EU) and the United States Department of Energy (US-DOE). Table 8 provides the key assumptions for the models discussed.

APERC [4] assumes the strongest increase in GDP while the WEO 98 assumes the lowest. As for world oil prices, APERC assumes a flat price, while EU [3], US-DOE [2] and WEO 98 all foresee future increases.

APERC projects energy demand to grow significantly faster, at an annual growth rate of 2.2% against 1.2% for WEO 98. This can be mainly attributed to the 50% higher GDP growth rate assumption made by APERC. There is a difference in the change in the market share of coal. While WEO 98 projects a market loss of close to 3 percentage points for coal, APERC expects an increase of close to 4 percentage points. The oil share in the power input mix declines more rapidly in WEO 98 than in the APERC projections, while WEO 98 projects higher market penetration rates for gas.

In terms of total energy demand, the IEA, US-DOE and EU models project a similar annual growth rate up to 2020 as shown in Table 9. A striking feature of the EU's projection is that gas demand is expected to show almost no growth. Both US-DOE and IEA project gas to be the most rapidly increasing fuel type. For oil, EU and US-DOE projections are higher than that of WEO 98. EU and WEO 98 project similar growth trends for transportation, 1.3% and 1.2% per annum, respectively. The main difference between the EU and WEO 98 oil figures is in the assessment of the power generation sector. While WEO 98 expects a declining trend of −0.4% per annum, the EU projects a doubling of oil in power generation, implying a growth rate of 2.6% up to 2020. For hydro and renewable energy sources, both organisations project significantly more conservative trends than does WEO 98.

Corresponding author. Fax: +33-1-4057-6659; email: fatih.birol@iea.org

¹ These regions are: OECD Europe, OECD North America and OECD Pacific grouped together as OECD; the Transition Economies (i.e., the countries of Central and Eastern Europe and of the former Soviet Union); China; East Asia, South Asia, Latin America, Africa and the Middle East, grouped as Rest of the World (ROW).

² The purchasing power parity (PPP) method converts goods and currencies into a common currency using a common set of prices, thereby obtaining real, comparable quantities. For the projection period, a constant relationship between PPP dollars and market exchange dollars has been assumed using 1995 data.

³ If biomass is included in the projections, then world energy demand is expected to increase from 9245 Mtoe in 1995 to 14,995 Mtoe in 2020.