Energy, Carbon Dioxide Emissions, and Economic Growth

Executive Summary
Fossil fuels account for 85 percent of total energy U.S. consumption. To reach the proposed Kyoto Protocol target, the United States would need to cut fossil fuel use by 15 percent and total energy use by 13 percent from current levels. This translates into a 1.8 percent decline in energy use per worker each year between now and 2010-which in turn would cut U.S. productivity growth by one-half and result in a standard of living 15 percent below what now is projected. Manufacturing, personal and commercial transportation, and the electric utility sectors would be especially hard-hit by the Kyoto Protocol, and everyone would feel its effect, even if fossil fuels were not rationed and energy prices did not suddenly increase.

Three reasons explain why the United States is the largest energy consumer among world economies. First, we have the world’s most energy-intensive capital stocks. Second, rising living standards and a growing population have generated large demands for energy. And third, energy remains relatively inexpensive in the United States.

Given these unique historical circumstances, and the fact that 99 percent of U.S. carbon dioxide (CO2) emissions comes from burning fossil fuels, the Kyoto Protocol to the Framework Convention on Climate Change would have a very significant impact on U.S. productivity growth and living standards. In December, 1997, the Clinton Administration, but not the U.S. Congress, endorsed the Protocol which, if ratified and implemented, would force the United States to reduce greenhouse gas (GHG) emissions in 2008-2012 to 7 percent below 1990 levels. Put in purely quantitative terms, U.S. fossil fuel consumption would need to decline from 79 quadrillion Btu (1996 consumption) to 67 quadrillion Btu (the projected 2010 Kyoto target). This paper looks in more detail at what such a reduction would mean to U.S. businesses and industries, workers and families.

Recent Trends in U.S. Energy Consumption

Total Energy Use
Total U.S. energy consumption of fossil fuels, nuclear, and hydroelectric power rose 2.6-fold from 33.992 quadrillion Btu in 1950 to 89.843 quadrillion Btu in 1996 (see Figure 1). Growth in consumption averaged 3.5 percent per year during 1950-1972, but only 0.8 percent per year from 1973-1996. There are two important reasons for the sharper growth during the 1950-1972 period: real energy prices declined by 30 percent during these years, and real GDP grew more rapidly during the period 1950-1972 than during the years 1973-1996. The decline in real energy prices from 1950 until 1972 combined with rapid growth in real GDP to promote strong growth in the demand for energy.

[Figure 1: Total and Per Capita Energy Consumption From Fossil Fuels, Hydro, and Nuclear Power in the United States, 1950-1996]

Source: DeGolyer and MacNaughton, Twentieth Century Petroleum Statistics, 1997(Dallas, Texas: DeGolyer and MacNaughton, 1997), p. 108.

This situation changed abruptly between 1973 and 1982. Following the initial OPEC embargo on oil exports, real energy prices jumped 55 percent in 1974, then rose more gradually each year through 1978. However, energy prices doubled again between 1979 and 1981 in response to the Iran-Iraq War and cutbacks in OPEC oil exports. Not surprisingly, total U.S. energy consumption dipped from 1973 until 1975, rose until 1979, then dropped sharply from 1979 until 1983. Notice that U.S. energy consumption totals in 1986 and 1973 are nearly identical (see Figure 1).

Energy prices began to drift downward after 1981, then plunged during 1986. As energy prices fell during and after 1986, total energy consumption grew by about 21 percent between 1986 and 1996.

Per Capita Energy Use

Energy use per capita increased from 223.8 million Btu in 1950 to 351.3 million Btu in 1973 (see Figure 1). This amounts to an average increase of 2 percent per year. But between 1973 and 1982, rising energy prices caused energy per capita to fall by 14 percent, from 351.3 million to 305.4 million Btu. Then, as real energy prices fell after 1986, energy per capita increased from 308.5 million to 338.7 million Btu, or by 10 percent.

These trends in total and per capita energy consumption show that energy use responds to changes in real energy prices. During periods of falling prices, such as 1950-1972 and 1986-1996, energy consumption increased. When prices rose between 1973 and 1982, energy use declined. But price and consumption patterns are hardly the full picture of what energy means to the U.S. economy. Technological progress, energy per worker, and capital per worker are all important determinants of productivity. And U.S. productivity growth is enhanced by growth in energy and capital per worker. Conversely, productivity declines when either energy use or capital per worker declines.

Trends in Productivity Growth

Why is commercial energy so vitally important in the United States? The reasons are straightforward. Factories must have energy from natural gas, refined oil products, coal, or electricity. Modern cars, trucks, buses, railroads, airplanes, and ships require gasoline, diesel fuel, jet fuel, or bunker fuel. Modern agricultural equipment cannot run without diesel fuel, and we rely on natural gas, heating oil, and electricity to maintain comfort in our homes, hospitals, and offices.

One way to understand the link between energy, physical capital formation, and U.S. productivity growth is to examine two distinct epochs between 1950 and 1984. Epoch 1 (1950-1972) represents years of falling real energy prices, steady growth in energy and utilized capital per worker hour, and rapid productivity growth of 2.6 percent per year. Epoch 2 (1973-1984) is a period of rising energy prices, declining energy use per worker hour, and a decline in productivity growth to only 1.0 percent per year.

Note that between 1950 and 1972, real GNP per hour (an economy-wide measure of labor productivity) increased by 2.6 percent per year (see Figure 2). Energy per hour also grew by 2.6 percent per year and utilized capital increased by 2.4 percent per year. The rapid 2.6 percent per year growth in productivity during Epoch 1 is attributable to balanced growth in energy and capital per hour and technological progress.

[Figure 2: Real GNP, Utilized Capital, and Energy Per Worker Hour, 1950-1984]

By contrast, sharply rising energy prices during Epoch 2 caused several abrupt changes. First, between 1973 and 1984, energy use declined by 1.3 percent annually. Utilized capital per hour continued to increase, but more slowly than during Epoch 1. The reduction in energy per hour combined with slower growth in capital per hour to restrict productivity growth during Epoch 2 to only 1.0 percent per year. Increasing energy and capital per hour play important roles in promoting productivity growth. To restrict energy per worker is to curtail productivity growth.

Fossil Fuels and the Kyoto Protocol Target

In 1996, the United States consumed 89.84 quadrillion Btu of energy from fossil fuels, hydroelectricity, and nuclear-generated electricity, and 92.83 quadrillion Btu from all sources (see Table 1). Of this total, 78.83 quadrillion Btu, or 85 percent, came from fossil fuels.

Table 1 U.S. Energy Consumption From Fossil Fuels, Renewables, and Nuclear Power (Quadrillion Btu)
Fossil Fuels
(including hydro)
Where we were in 1996 78.83 6.83 7.17 92.83
Where we were in 1990 71.98 6.16 6.02 84.16
Where Kyoto Protocol would put us in 2010 66.94 6.83 7.17 80.94
Reduction from 1996 in energy consumption due to Kyoto Protocol -11.89 0 0 -11.89
Sources: Energy from all renewable resources is from the U.S. Energy Information Administration, reported in Statistical Abstract of the United States, 1997, p. 599. The estimate of 6.83 quads is for 1995, not 1996.

Energy estimates from fossil fuels and nuclear power are from the U.S. Energy Information Administration, reported in DeGolyer and MacNaughton, Twentieth Century Petroleum Statistics, 1997 (Dallas, Texas: DeGolyer and MacNaughton, 1997), pp. 109-112. Fossil fuel energy in 2010 under the Kyoto Protocol is 7 percent less than that in 1990, i.e., 71.98 x .93 = 66.94 quadrillion Btu.


In 1990, the United States consumed 84.16 quadrillion Btu of energy from all sources. Fossil fuels made up 71.98 quadrillion Btu, or 85.5 percent of the total. The Kyoto Protocol calls for the United States to reduce GHG emissions to 93 percent of the 1990 rate sometime during the years 2008-2012. I will use 2010 as the target year. Reducing GHG emissions is not the same as reducing CO2 emissions to 93 percent of the 1990 rate because CO2 is not the only GHG. But it is by far the most important as a potential source of global warming. The Kyoto Protocol mentions trade-offs between different GHG reductions, but they are too vague to be analytically useful.

Reducing CO2 With a Fixed Fossil Fuel Mix

My analysis of the cutback in fossil fuels treats the mix of natural gas, oil, and coal as fixed. In short, I ignore any prospective substitution among fuels. Doing so allows me to sidestep an important question: How would the prospective cutback be achieved? One widely discussed proposal is to impose a tax on the carbon contained in fossil fuels. Their carbon content and emissions rates differ quite a lot. Coal is the most carbon-intensive fuel, natural gas the least. Because the carbon (and CO2) emissions rate per unit of energy from coal is nearly twice as great as that from natural gas, a carbon tax would increase the price of coal the most, the price of natural gas the least.

In any case, very little substitution could occur within a year or so, because most capital equipment designed to burn one fuel cannot easily be adapted to use another. How much fuel substitution would occur over several years depends upon several things: (1) the tax-induced changes in relative fuel costs; (2) the rate of retirement of coal-using capital; and (3) rates of investment in capital designed to burn other fuels. The important point is that a carbon tax would provoke long-run substitution of less carbon-intensive for more carbon-intensive fuels. Given so many uncertainties surrounding carbon taxes, tradable emissions permits, and the Kyoto Protocol itself, there are many good reasons to set aside for now how the Protocol would affect the mix of fuels in the United States.

Assuming that the ratios in which petroleum, natural gas, and coal are consumed remain as they were in 1996, then reducing CO2 from fossil fuels would require that the 1996 fuel mix be scaled back to get aggregate Btu from fossil fuels in 2010 equal to 93 percent of aggregate Btu from fossil fuels in 1990, or .93 x 71.98 quadrillion Btu = 66.94 quadrillion Btu. This amounts to a reduction from 1996 of 78.83 – 66.94 = 11.89 quadrillion Btu. So, to meet Kyoto Protocol targets would require that U.S. fossil fuel consumption be cut back to (66.94 / 78.83) = .849, or to 84.9 percent of actual consumption in 1996.

Total energy consumption need not be cut back so much, of course, because renewable energy and nuclear energy do not produce CO2. In theory, renewable and nuclear energy could grow to fill the gap created by reduced fossil fuels. But in practice, this energy loss of 11.89 quadrillion Btu cannot be filled in 12 years by expanding either renewable or nuclear energy.

There are two compelling reasons why this is so. First, U.S. nuclear-generation capacity has not increased since 1990. We had 110 nuclear power plants in 1996, the same number as in 1989. No new nuclear plants have been licensed since 1978. Nuclear power in the United States cannot be increased without new plants. Even if licensing were to be resumed immediately, and it almost surely will not, nuclear energy could be increased very little, if at all, by 2010. For the next 12 years, it seems best to project U.S. nuclear energy as either fixed or declining.

Second, for the next 12 years, expansion of renewable energy is strictly limited. Hydroelectric power is the same in 1996 as it was in 1990, and only marginally greater than in 1980. The estimated remaining undeveloped hydropower in the United States is a drop in the bucket. Further development of hydroelectricity faces strong opposition from environmental groups. Since hydroelectricity is half of all renewable energy, and renewable energy amounted to only 7.5 percent of all U.S. energy in 1996, we can regard renewable energy as practically fixed for the next 12 years. Even if non-hydro renewable energy were to double in the next 12 years (and it cannot do so), renewables would not begin to offset the 12 quadrillion Btu loss attributable to the reduction in fossil fuels.

Fossil Fuel Use Reduced by 15 Percent

The Kyoto Protocol requires no reduction in renewable or nuclear energy. So to cut CO2 emissions to 93 percent of the 1990 level commits the United States to reduce fossil fuels from 78.83 quads in 1996 to 66.94 quads in 2010. This amounts to a 15 percent cutback in fossil fuels from 1996 (-11.89 / 78.83 = -.151, or -15 percent). It also amounts to a 13 percent cutback in overall energy between now and 2010 (-11.89 / 92.83) = -.128, or -13 percent).

The United States has seldom experienced such a large energy reduction. Two comparisons are suggestive. Between 1973 and 1975, the period of the first OPEC oil embargo, energy consumption fell from 74.2 to 70.5 quads, but rebounded sharply between 1975 and 1979. Then from 1979 to 1983, a period of rising energy prices caused by lower OPEC production, energy consumption fell from 78.7 to 70.4 quads, but rebounded to 79.9 quads by 1988. The decade 1973-1983 was marked by energy shocks and economic stagnation: Maddison (1995) shows that real GDP per capita increased by only 1.1 percent per year, and our calculations show that real GNP per hour increased scarcely 0.8 percent per year. We recall the decade 1973-1983 as a period of slow growth in productivity and living standards.

Energy Use Per Worker Down by 28 Percent

The Kyoto Protocol would require a 15 percent cutback in fossil fuels and a 13 percent reduction in total energy by 2010. It would require an even larger cutback in energy per worker because our population and labor force will continue to grow. Energy per worker and capital per worker are key to productivity growth.

To quantify these reductions, begin with U.S. Census Bureau population estimates for 1990 (249.4 million) and 1996 (265.3 million), and project the population in 2010 to be 305 million. Next, divide the actual energy consumption in 1990 and 1996 (shown in Table 1) by the actual population in those years to obtain energy consumption per capita in 1990 (320 million Btu) and 1996 (349.9 million Btu; see Table 2). Then find the per capita energy use available under the Kyoto Protocol (265.4 million Btu) by dividing energy consumption in 2010 (Table 2) by the projected U.S. population of 305 million people in 2010. All other things being equal, and factoring in labor force growth, the Kyoto Protocol would require a 28 percent reduction in energy use (84.5 million Btu) per worker, compared to 1996. To reach this target smoothly, energy use per worker would need to decline 1.8 percent each year between 1996 and 2010./TD>

Table 2 U.S. Per Capita Energy Consumption From Fossil Fuels, Renewables, and Nuclear Power (Million Btu)
Fossil Fuels
(including hydro)
Where we were in 1996 (actual population of
265.3 million)
297.1 25.8 27.0 349.9
Where we were in 1990 (actual population of 249.4 million) 271.2 24.7 24.1 320.0
Where Kyoto Protocol would put us in 2010 (projected population of 305 million) 219.5 22.4 23.5 265.4
Reduction from 1996 in
energy consumption due to Kyoto Protocol and population
-77.6 -3.4 -3.5 -84.5
Sources: Energy from all renewable resources is from the U.S. Energy Information Administration, reported in Statistical Abstract of the United States, 1997, p. 599.

Energy estimates from fossil fuels and nuclear power are from the U.S. Energy Information Administration, reported in DeGolyer and MacNaughton, Twentieth Century Petroleum Statistics, 1997 (Dallas, Texas: DeGolyer and MacNaughton, 1997), pp. 109-112.

Population in 1990 and 1996 is from the U.S. Census Bureau, reported in DeGolyer and MacNaughton, 1997, page 106.

Projected population in 2010 is based on 1 percent net annual growth for the years 1996-2010.


Productivity Growth Declines

To project annual productivity growth from 1996 until 2010 with no restrictions placed on energy, I assume that utilized capital per worker increases at 2.0 percent annually (as it did from 1950 until 1984) and that energy per worker increases at 1.3 percent annually (as it did from 1950 until 1984). Substituting these numbers into the equation below gives a “Business as Usual” annual productivity growth rate of 2.0 percent per year:


(1)  Annual % change in real GDP per worker =
.013 + .142 (.020) + .289 (.013) =
.020, or 2.0 percent per year.

The reduction in energy per worker of 1.8 percent implied by the Kyoto Protocol would reduce annual productivity growth from 2.0 percent to 1.1 percent, or by 0.9 percent per year. This projection of 1.1 percent productivity growth caused by declining energy per worker fits with the facts from 1973 until 1983: Energy per worker hour fell from 456 thousand Btu to 391 thousand Btu, and real GNP per hour increased by 0.8 percent per year.

I should stress several things about this projection. First, it has no energy shocks, but instead a steady annual reduction in energy per worker. Second, it retains the assumption of steady total factor productivity growth of 1.3 percent annually. Some might think this too optimistic under the 1.8 percent annual reduction in energy per worker. If a lot of energy-intensive capital must be scrapped, they are probably right. Third, it retains the assumption that utilized capital per worker continues to grow at 2.0 percent per year. If much energy-intensive capital must be scrapped, the net utilized capital stock may actually decrease. If so, gross investment in new energy-efficient capital will have to be much greater than it is now. Finally, it implies that Btu of energy per unit of capital declines quite rapidly at 3.8 percent per year. Such a strong decrease would require much higher gross investment in energy-efficient capital stocks than we have today.

My projections show that a shift from 1.3 percent annual growth in energy per worker (“Business as Usual”) to a 1.8 percent annual decrease (required by the Kyoto Protocol) would cut productivity growth from 2.0 percent to 1.1 percent. If the labor force grows by 1 percent (as I assume), then meeting the Kyoto Protocol would cut the U.S. GDP growth rate from 3.0 percent to 2.1 percent.

The differences between projected productivity growth of 1.1 percent under the Kyoto Protocol and 2.0 percent without it imply major differences in U.S. living standards by 2010. With productivity growth of 1.1 percent per year, U.S. living standards would be 17 percent higher in 2010 than in 1996. But with productivity growth of 2.0 percent per year, living standards would be 32 percent higher in 2010. The central conclusion is this: A 1.8 percent annual reduction in energy per capita that would be our commitment according to the Kyoto Protocol would lead to living standards in 2010 that are 15 percent lower than they would be with no restrictions on energy usage.


The 1.8 percent annual decrease in energy per worker would reduce the annual growth in real GDP per worker from 2.0 percent to 1.1 percent and reduce U.S. living standards some 15 percent below projected levels in 2010. This projected reduction in productivity growth is probably conservative. The main reason is that it does not account for the replacement of energy-intensive capital by the more energy-efficient capital that the energy cutback would call for. Replacing capital at the rate required to meet the Kyoto Protocol’s first budget period would place an enormous strain on private and government saving, and would require a much higher ratio of gross investment to GDP than we have today. Moreover, I have assumed that fossil fuel will not be rationed to meet the Kyoto target and that no energy price shocks will occur in the marketplace. This, of course, was not the case during fuel shortages in the 1970s and early 1980s, so my projections about the likely impact of the Kyoto Protocol on the U.S. economy, productivity growth, and lifestyles are probably quite conservative.


  • Bureau of the Census. 1997. Statistical Abstract of the United States: 1997.Washington, DC: Government Printing Office.
  • DeGolyer and MacNaughton. 1997. Twentieth Century Petroleum Statistics, 1997. Dallas, Texas: DeGolyer and MacNaughton.
  • Maddison, Angus. 1995. Monitoring the World Economy, 1820-1992. Paris: Organization for Economic Cooperation and Development.


This paper was prepared for the September 23, 1998, policy conference sponsored by the ACCF Center for Policy Research, and will be published in the Center’s forthcoming book, Climate Change Policy: Practical Strategies to Promote Economic Growth and Environmental Quality.