EFFICIENCY OF ENERGY USE

Although energy efficiency was already heavily emphasized in the 1970s as a key strategy for energy security, more recently it has also been recognized as one of the most cost-effective strategies for reducing environmentally harmful emissions. Energy efficiency is more than just a resource option such as choosing between coal, oil, or natural gas. It curbs demand rather than increasing supply, and thus provides additional economic value by preserving the resource base and reducing pollution.

For specific applications, we can calculate the ratio of the measure of the goods or services provided to the energy input required. For example, in the transportation sector, energy efficiency is based on miles per gallon for personal vehicles, seat-miles per gallon for mass transit, and ton-miles per gallon for freight transportation.

For the entire economy, with its countless services and inputs, economists usually define the "service" or economics efficiency as the entire GDP (Gross Domestic Product) divided by E, the annual total primary energy used: Economic Efficiency = GDP/E.

Economists also track the reciprocal, E/GDP, which is called energy intensity. For example, the energy intensity of the United States in 1998 was 91 quads/$8.5 trillion in 1996 dollars (1996$), which divides out to be 10,700 Btu/$. (Note: 1 "quad" = 1 Q = 1 quadrillion Btu = 10¹⁵ Btu.)

Measuring energy efficiency gains for the entire economy is not a precise science since the population continues to expand, new technologies continue to be introduced, and there is great variability in the behavior of individuals using technology. Nevertheless, trends in economic effiency and energy intensity best reflect the impact of energy efficiency improvements.

ORIGINS OF ENERGY EFFICIENCY

The increased availability of energy fueled the Industrial Revolution. The United States became the world's largest oil producer, and the new fossil fuels were abundant and modestly priced. A technology's energy efficiency was not a key part of capital investment decisions. Energy-efficient technology as a priority ranked well behind improved performance.

Energy intensity declined from 60,000 Btu/$ (1992$) in the 1850s to 13,000 units in 1995. There were rapid drops in the 1860s as the switch was made from wood to more efficient coal, and, starting in the 1920s, as the switch was made to even more efficient oil and gasoline. The 1973 OPEC oil embargo and the next eleven years of rising energy prices triggered the final dip. The overall drop by a factor of 4.6 in 145 years corresponds to a steady annualized efficiency gain of 1.1 percent.

Before the OPEC embargo, there was no Department of Energy, and energy efficiency was not considered to be a government responsibility. Other aspects of energy were understood to be appropriate for government support. For example, research and development (R&D) on futuristic power supply technologies such as fission nd fusion was funded by the Atomic Energy Commission. From fixed year (FY)1948 through FY1972, in 1999 constant dollars, the federal government spent about $22.4 billion for nuclear (fission and fusion) energy R&D and about $5.1 billion for fossil energy R&D. The government also had a role in electrification as an economic development strategy. The entire rural electrification effort, including the federally subsidized Power Marketing Administration is still a major government program today. But it took an OPEC embargo to convince Americans to create a Department of Energy (DOE) in 1974 and to use public funds for efficiency research and development.

RECENT TRENDS IN U.S. ENERGY INTENSITY

The relative lack of importance of energy prices changed dramatically with the OPEC oil embargo. Even though energy prices were still a small fraction of total costs, people and businesses began to make energy-efficient capital-investment decisions in expectation of higher prices.

From 1974 through 1992, Congress established several complementary energy-efficiency and energy-conservation programs. By the 1980s, this concern over finite resources had dissipated as higher prices encouraged greater innovation in efficiency and in resource recovery. Energy efficiency had become a cost-saving, "demand-side management" tool that helped to avoid expensive power plant construction. The DOE's 1995 report, Energy Conservation Trends, states that energy efficiency and conservation activities from 1973 through 1991 curbed the pre-1973 growth trend in primary energy use by about 18 Q, an 18 percent reduction.

By the late 1980s, concerns over air pollution began to play a role in the government rationale for energy efficiency, which in the 1990s was followed by concern over global warming. Neither concern had a strong impact on energy use. Since 1985, national energy use has climbed about 20 Q, reaching a record high of 92 Q in 1999. From FY1980 through FY1999, the DOE spent $7 billion on energy efficiency R&D, which accounted for about 10 percent of all energy supply R&D.

E/GDP experienced a steep decline during the eleven OPEC years (1974 through 1985) and a recent equally steep decline starting in 1997. The latter drops may be associated with the rapid growth of the U.S. economy and with the explosive growth in information technology and the Internet. From 1960 through 1973, energy prices were low and there was almost no improvement in E/GDP. Similarly, after the collapse of OPEC in 1985, prices were again low, energy policy wavered, and E/GDP leveled off. The overall drop from 18 in 1973 (the year before the embargo) to 10.5 in 1998 is a drop of 57 percent, and corresponds to a steady gain of 2.2 percent/year for 26 successive years.

Thus, improved energy efficiency can be credited with energy savings of $232 billion in 1999. The arithmetic is as follows. A drop to 57 percent corresponds to a savings of 43 percent; but roughly one-third of the gain (about 15%) came from structural change as we switched from a smokestack to a service economy, and only two-thirds (about 30%) came from a true increase in efficiency. So pure efficiency has reduced our energy intensity only to 70 percent (not 57%). Had our efficiency stayed frozen at its 1973 value, we would now use more energy by the factor 1/0.7 = 1.43, that is, we would use 43 percent more energy for every dollar of GNP. Our 1999 energy bill was $540 billion, and would have been larger by 43 percent, which is $232 billion. The fraction that can be attributed to government intervention in the marketplace is highly debatable. Clearly, the marketplace would have made the energy efficiency improvements anyway; yet, a considerable portion would not have taken place without government intervention.

It is interesting to compare this huge annual saving of $232 billion with two other 1999 expenditures. The total non-military discretionary federal budget was $300 billion; therefore, efficiency savings pay for three-fourths of our entire civilian discretionary budget. Efficiency savings also equate to a large percentage of the U.S. Social Security budget, which in 1999 was $392 billion.

As technology develops steadily, there follows a corresponding decline of E/GDP, averaging about 1 percent/year. This can be accelerated in the marketplace by new fuels, new technologies, and innovations in existing technology. Government intervention—such as efficiency labels, performance standards for buildings and equipment, tax incentives, utility policy, and voluntary agreements with industry—which is usually implemented during periods of rising energy prices—can further accelerate the decline in E/GDP.

INTERNATIONAL COMPARISONS OF ENERGY INTENSITY

The E/GDP for the United States has sloped steadily downward from 18,000 to 11,000 Btu/$. Europe and Japan are typically only half as energy-intensive as the United States. An explanation is that, during their development, Western Europe and Japan were petroleum-poor compared to the United States, so energy use was perceived to imply imports (and risk of supply disruption) and trade deficits. Thus, they adopted tax policies to conserve energy. The United States took the opposite path; to stimulate economic growth, domestic oil and gas production was subsidized.

Among E/GDPs for developing countries or regions, the most notable is that of China, reaching 110,000 to 120,000 Btu/$ until 1976, but then declining steadily 5.2 percent/year for 21 years to 40,000 in 1997. This two-thirds drop shows the striking potential savings for other developing economies. The former Soviet Union (FSU) had a steady but inefficient economy until 1989. After that the FSU's rise in E/GDP is mainly because of the collapse of GDP. Eastern Europe comes next. It started off indistinguishably from the Soviet Union, made small improvements until 1989, and then rapid improvements as it adopted market economies. It is expected that the FSU curve will also soon turn down as its GDP picks up. Although India's efficiency trends are not currently in the right direction, developing countries, and particularly the FSU, have a high potential for cost-effective efficiency gains. Well below India come the industrialized countries, with the United States at the top and Japan at the bottom.

ENERGY SAVINGS IN THE BUILDINGS SECTOR

Energy is used in buildings to provide a variety of services such as lighting, space conditioning, refrigeration, hot water, and electronics. In the United States, building energy consumption accounts for slightly more than one-third of total primary energy consumption. Percentages reported for energy consumption and related carbon emissions in all four sectors are based on the Energy Information Administration's Annual Energy Outlook 2000, DOE/EIA-0383(2000), December 1999. By 2010, significant changes are expected to occur that will affect how buildings are constructed, the materials and systems used to build them, and the way in which buildings are maintained and used. A wide array of technologies can reduce energy use in residential and commercial buildings. Using sensors and controls to better manage building energy use, and improving building design and construction materials to maximize the thermal resistance of the building shell can also significantly reduce building energy requirements. For example, a cool white roof can reduce air-conditioning energy use by 20 percent.

Appliances have shown very dramatic improvements in energy efficiency, and perhaps the most impressive efficiency gains have come in improving refrigerators. In what follows, we show that refrigerators' efficiency gains are due to the interplay of regulatory and technological advancement. Two energy regulatory innovations (appliance labels, soon followed by standards) and a major technological innovation (blown-in foam insulation) led to the change from an annual energy use growth of 7 percent/year to a drop of 5 percent/year.

In the 27 years between the 1974 peak annual usage of 1,800 kWh and the 2001 federal standard of 450 kWh, refrigerator energy use dropped to one-quarter of its former use, even as the average volume grew from 18 cu. ft. to 20 cu. ft. This corresponds to a compound annual efficiency gain of 5.1 percent. As for economic savings, by the time 150 million refrigerators have reached year-2001 efficiency, compared to 1974, they will save 200 billion kWh/year, which corresponds to the output of forty huge (1 GW) power plants, and to one-third of the nuclear electricity supplied last year in the United States. Consumers will save annually $16 billion annually in electric bills, but their net savings will be only $10–11 billion, because there is a cost premium for the improved refrigerator (typically repaid by bill savings in three years). This $16 billion annual electricity saving from refrigerators alone roughly matches the entire $17 billion wholesale annual value of all United States nuclear electricity.

This surprising equality arises because an efficient appliance saves "expensive" electricity at the meter, at an average retail price of 8 cents/kWh; whereas one kWh of new wholesale supply is worth only 2–3 cents at the power plant. Thus, even if electricity from some future new remote power plant is "too cheap to meter," it still must be transmitted, distributed, and managed for 5–6 cents/kWh. It is impossible to disentangle the contribution of standards and of accelerated improvement in technology, but clearly the combination has served society well.

ENERGY SAVINGS IN THE INDUSTRIAL SECTOR

The industrial sector is extraordinarily complex and heterogeneous. It includes all manufacturing, as well as agriculture, mining, and construction. In the United States, industrial energy consumption accounts for slightly more than one-third of total primary energy consumption. Recent data show nonmanufacturing industries such as agriculture and construction have maintained their energy use growth rate while that for manufacturing has dropped.

Still, the manufacturing sub-sector accounts for about 70 percent of industrial-sector energy consumption. Nonenergy-intensive manufacturing accounts for an increasing share of energy use; for example, electronic equipment is expected to have a growth rate twice that of the manufacturing sector as a whole. The most energy-intensive (in terms of energy used per dollar of output) manufacturers are iron and steel, pulp and paper, petroleum refining, chemicals, and cement; together, these industries account for about half of the primary energy consumed in the industrial sector.

Of the end-use sectors, the industrial sector—especially in its more energy intensive industries—has shown the greatest and the fastest energy efficiency improvements. For example, over the past quarter century, the U.S. steel industry has reduced its energy intensity by nearly 50 percent; the cement industry has improved its fuel efficiency nearly 30 percent since 1975 although, since 1986, the energy intensity improvements have slowed somewhat. The energy use per pound of product in the chemicals industry has fallen at an average of 2 percent per year, and its energy efficiency continued to increase during periods when energy price was stable or falling (though less steeply when the price was falling). By using more of its former waste products for energy, the pulp and paper industry increased its purchased fuel efficiency by nearly 45 percent from 1972 to 1994.

Clearly, the more recent industrial energy efficiency gains are due more to technological progress than to energy prices. Unlike the buildings and transportation sectors, industry has adopted some supply-side energy-efficient technologies that reduce emissions without necessarily reducing energy demand. These include more efficient use of byproduct fuels and retrofitting boilers for combined heat and power. On the electricity demand side, some generic improvements, such as high-efficiency motors and advanced motor system drives and controls, have applications in almost all types of industry.

Other energy-efficient technologies are more sector-specific. For example, in the future, the steel industry, even with increases in recycling, will need to make some steel from ore. A new cokeless steelmaking process could cut energy use 30 percent relative to a blast furnace by going directly from solid ore to steel. This "smelt reduction" technique could also increase the industry's productivity, as its investment costs and operating costs are much lower. There are many reduction opportunities of this order of magnitude in various industries, but the single biggest "bang for the buck" is in more efficient heat and power systems.

Combined Heat and Power (CHP) systems, also called "cogeneration" systems, generate electricity (or mechanical energy) and heat simultaneously at the point of use. Figure 1 shows that in 1994, manufacturers used 7.4 quads to generate electricity and, together with the on-site steam produced from separate boilers, required 16.3 quads of fuel, for a system thermal efficiency of 46.5 percent. If produced jointly as CHP at 85 percent efficiency (clearly achievable based on the previous figure), the total fuel requirements would be only 8.9 quads, nearly 50 percent less. Replacing much of industrial Separate Heat and Power (SHP) with CHP by 2010 is not so far-fetched. According to one source (Kaarsberg and Roop, 1999), more than 75 percent of the industrial thermal capacity installed today will be retired by 2010.

ENERGY SAVINGS IN THE TRANSPORTATION SECTOR

Transportation accounts for about one-quarter of total U.S. primary energy consumption. Since 1986, average new car horsepower has increased nearly 40 percent. The average miles per gallon (mpg) of new light-duty vehicles, new cars, and light trucks combined has not changed significantly since 1982. In the absence of new efficiency standards, carmakers' technical improvements respond to consumer demands for roomy, powerful vehicles. Between 1986 and 1997, the average fuel economy for new passenger cars increased by less than 2 percent (from 28.2 to 28.6 mpg), while the horsepower (hp) per weight increased by 27 percent (from 3.89 to 4.95 hp/100 lb), and weight further grew by 9 percent. The fuel economy of the entire fleet (including a growing fraction of light trucks) decreased of 6 percent (0.5% /year) over this period. Figure 2 shows our calculation of what the fleet fuel economy could have been if carmakers had focused on reducing fuel use rather than increasing power, weight, and size.

PARTNERSHIP FOR A NEW GENERATION OF VEHICLES (PNGV)

The PNGV is a government-industry (General Motors, Ford, DaimlerChrysler) research partnership. One of the most highly publicized PNGV goals is to triple the fuel efficiency of a car (with a prototype by 2004) while preserving safety, performance, amenities, recyclability, and holding down costs. As a result of PNGV, federal government R&D in advanced automotive technologies has been reorganized and redirected toward this ambitious goal. There are PNGV programs in advanced materials; electric drives, including power electronics; high-power energy storage devices; fuel cells; and high efficiency low-emission diesel engines.

A key element of the PNGV is the development of an advanced hybrid-electric vehicle. Energy losses from conventional engine idling or running at part load are eliminated in hybrid vehicles-which are available today. Efficiency is doubled with such hybrid propulsion systems. One-hundred units of fuel needed in today's new car (averaging 28.6-mpg) produces the same amount of drive power as 50 units needed by the electric hybrid. Even today's relatively efficient gasoline spark-ignition internal combustion (IC) engine loses 84 units of energy per 100 units of fuel in. The most easily reduced of these losses are the "standby" losses, which account for 11 percent on average (or up to 20% under increasingly typical congested conditions). These losses occur when the engine is either idling or running at far less than 100 percent load. Standby losses are so high because the engine is oversized to allow for acceleration and therefore runs almost entirely at less-efficient part load. Other losses, in descending order, are exhaust, radiator, engine friction and pumping losses, and accessories (electrical system, pumps, etc., but not including major seasonal loads such as air conditioning). Once the power is delivered to the drive train, it is finally completely dissipated in braking and in rolling and wind resistance.

The 2x electric hybrid suffers no standby engine losses. It features a less powerful, more fuel-efficient IC engine/generator running at full load (or off) and uses a battery-powered electric motor to boost acceleration. Thus, the IC engine is not oversized. It has only two modes: maximally efficient full load, or "off." Engine power typically is delivered not to a drive shaft, but to a generator, on to a storage battery, and then to electric motors on the wheels. While the engine is off, the battery powers these motors. When the engine is on, the battery is either being recharged or is boosting acceleration. The hybrid configuration also helps to reduce driving losses. Instead of friction-braking, an electric hybrid uses its electric motor as a generator to recapture braking energy and charge the battery. Thus, braking losses are reduced by 70 percent. They would only be eliminated if the batteries and generator were 100 percent efficient and all four wheels could recapture the braking energy.

Energy losses also occur as the car is propelled. The 16 units of power are dissipated in rolling resistance and aerodynamic drag (and in today's cars, in friction-braking). The next challenge to reach 3x is to reduce the weight and rolling losses with lighter materials and reduce drag with sleeker designs. A major obstacle to reaching the PNGV affordability goal is the high cost of advanced, lightweight body and tire materials. In such a path, about half the savings are due to the propulsion system and half to the improved (lighter-weight, more aerodynamic) envelope.

This is only one of the many "paths" to achieve 3x efficiency. Other ways to eliminate standby losses being investigated by the PNGV program include advanced diesels, direct injection stratified charge gasoline engines, and fuel cells. With these options, the engine's efficiency could double without going to a hybrid configuration.

It is too soon to judge the PNGV program. Surprisingly, Honda and Toyota were the first companies to introduce 3x efficiency hybrids into the U.S. marketplace in 1999 and 2000, respectively. These two companies were not part of the PNGV program, and thus the early introduction shows that the marketplace could efficiently develop and market efficient vehicle technology without the aid of government. Whether the PNGV participants can catch up and market far superior hybrid vehicles is yet to be seen.

AN OPTIMISTIC CONCLUSION

Above, we said that (apart from crises like the eleven OPEC years) energy intensity E/GDP falls about 1 percent year (E is inversely proportional to energy efficiency, η). Here we are concerned with trends in United States, and later, world energy use, so we write: E = E/GDP × GDP =Const/η× GDP Then to stabilize E, η must rise annually not by merely 1 percent, but fast enough to cancel our desired annual growth in GDP (or gross world product GWP), that is, about 3 percent. We have discussed six examples of why this is achievable, given sufficient motivation to do so:

1. For the last three years, in the United States, the annual growth in η has averaged not 1 percent, but 3.5 percent; that is, while GDP has surged nearly 4 percent/year, energy use has leveled off. It is still unclear how much of this gain is a real trend from increased productivity and the explosive growth of information technology and the Internet (particularly business-to-business e-commerce). But if even part of this gain continues, it is very good news for reducing carbon emissions.
2. During the eleven OPEC "crisis" years, it is well known that η grew 3 percent/year, of which 1 percent was "structural" (moving from a "smokestack" to a service economy) and 2 percent was a pure efficiency gain. This annual 2 percent is measured for our whole stock of energy using equipment, most of which has a service life longer than the 11-year "experiment." Thus, cars last 12 years, refrigerators 15, buildings 50, and so on. So this gain in the stock must lag the gain in new products, or (re-worded) the rate of improvement of new products must lead that of the stock. Arthur Rosenfeld and David Bassett (1999) have crudely estimated this lead/lag correction, and find that new products improved 5 percent/year.
3. Our refrigerator discussion showed that under appliance standards, refrigerator energy use has been dropping more than 5 percent/year for 27 successive years. Yet the payback time for the improving technology has stayed at two to five years. This suggests that significant steady gains can be kept up for a very long time.
4. During the OPEC years, auto fuel economy improved 7 percent/year. After correcting for the 40 percent increase in power, we see an adjusted gain from 1975 through 1997 of 4 percent/year. If the PNGV 3x car at 80 mpg is a significant fraction of the new car fleet by 2010, this rate of improvement will have been sustained for 35 years.
5. Combined Heat and Power (CHP) is generally 1.5-2 times more efficient than separate heat and power. It grew nearly five-fold in the United States in the years between the enactment of CHP incentives with the passage of the Public Utilities Regulatory Policy Act of 1978 (PURPA) and today. During some of that period it grew at more than 15 percent per year and now accounts for 9 percent of electricity generation.

THE WORLD'S NEED FOR ENERGY IN 2100

Next we estimate the world's need for energy in 2100 (E_w), under three scenarios, where the symbol α will denote the annual gain in energy efficiency η:

1. "BAU" (Business as Usual): α = 1 percent/year, its historic non-crisis rate.
2. "No Regrets": α = 2 percent/year; that is, the world runs scared of climate change.
3. "De-Materialization": α = 3 percent/year; a bit less than the actual rate for the United States for 1997 to 2000.

For this brief discussion, we factor world energy in 2100 as where Pop.₂₁₀₀ ≈ 10 billion and α = –(d/dt)(E/GDP) = 1 to 3 percent per year. Population by 2100 will probably level off at slightly under 10 billion, and Watts_p(2000)/capita is the rate of primary energy use today, considered a satisfactory goal by the majority of people in developing countries today. For Watts_p(2000)/capita we propose 5 kW (that is, 5 kW-years of energy for the year 2000), corresponding to that of Western Europe today. Thus, we assert that a poor African or Indian today would happily aspire to a year-2100 standard of living, health, transportation, and so on equal to that of Germany or Scandinavia today (even if they have few SUVs).

We also note a 1985 study (by Goldemberg et al.) showing that the best then-available technology could yield a Western European lifestyle at only 1 watt of primary energy per capita.

The exponential factor arises because, in 100 years, this 5 kW/capita will drop: Next, we note that world primary power today is slightly over 10 TW, so we can write:

Table 1 shows how many 2000 Worlds of power

we must construct by 2100. Many authors ignore the sensitive dependence on α and discuss the need for a huge program of technology development and construction. Our view is that in any given year it is cheaper and easier to improve efficiency by 2 percent than to add 2 percent to the world's entire energy supply. Since this can be done purely by investments that produce net (life-cycle) savings, we call it a "No Regrets" scenario. The third line of Table 1 corresponds roughly to the current α for the United States. If this were to continue, and spread globally, gross world product can grow at a healthy annual 3 percent, and energy use could level off at today's rate. If we raised α only to 2 percent we would need to grow supply by only 1 percent/year.

Of course, we in the present developed world will not give up "our" present 7–8 TW, but Equation 3 and Table 1 show that under our "No Regrets" scenario, 10 billion people in developing countries need add only 0.5 new "worlds" of energy supply to provide them with an attractive current Western European standard of living. This is what we labeled above as an "Optimistic Conclusion," but it is a continuous and enduring challenge to the inventiveness of technologists and policy-makers of the twenty-first century.

Arthur H. RosenfeldTina M. KaarsbergJoseph J. Romm

See also: Cogeneration; Hybrid Vehicles.

BIBLIOGRAPHY

Goldemberg, J.; Johansson, T. B.; Reddy, A. K. N.; and Williams, R. H. (1985). "Basic Needs and Much More in One Kilowatt per Capita." Ambio 14(4–5):190–200.

Kaarsberg, T. M., and Roop, J. M. (1999). "Combined Heat and Power in Industry: How Much Carbon and Energy can Manufacturers Save?" IEEE Aerospace and Electronic Systems Magazine, pp. 7-11, January

Rosenfeld, A. H., and Bassett, D. (1999). "The Dependence of Annual Energy Efficiency Improvement on Price and Policy." International Workshop on Technologies to Reduce Greenhouse Gas Emissions. Crystal City, VA, May 1999, Plenary IV. Paris, France: International Energy Agency. <http://www.IEA.org/workshop/engecon/>.

U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy. (1997). Scenarios of U.S. Carbon Reductions Potential Impacts of Energy Technologies by 2010 and Beyond. <www.ornl.gov/ORNL/Energy_Eff/labweb.htm>.