ASSESSING CARBON EMISSION CONTROL STRATEGIES:

T H E C A S E OF C H I N A

W I L L I A M U. C H A N D L E R
Battelle, Pacific Northwest Laboratories, 370 L'Enj&nt Promenade, S.W. Washington, D.C. 20024,
U.S.A.

Abstract. Scientists are now being asked to recommend measures to reduce the
risks of climatic change due to anthropogenic greenhouse gases. Considerably
less effort, however, has been allotted to understanding the efficacy of controlling
these gases than to their effects. This paper briefly describes and applies an
energy-economic model to assess the effectiveness of carbon dioxide control
policies that theoretically could be enacted in China, a large, developing
nation with an energy inefficient and 'carbon-intensive' economy. The paper
also assesses the effectiveness of similar international efforts, as well as the
effect of each initiative on Chinese income levels. Carbon dioxide control
measures are contained in scenarios drawn to the year 2075 and include family
planning, fossil fuel taxes, mandatory or technical energy efficiency improve-
ments, and a combination of these.
The results suggest, not surprisingly, that no nation alone, not even China,
can decisively affect the global CO2 problem. More importantly, however,
the potential for energy efficiency improvements in China is found to be both
very large and economically attractive. Scenario analysis suggests that energy
effficiency measures could both reduce carbon emissions significantly and
increase Chinese per capita incomes. Similar conclusions are drawn regarding
worldwide energy-efficiency measures. Thus, appropriate public policy measures
to capture the existing energy-efficiency potential might both reduce the risk
of climatic change and improve economic standards of living.

I. Introduction

Can and should policy-makers act to reduce the risk of climatic change due to
anthropogenic greenhouse gases? As scientists have grappled with this question,
a schism has appeared between what has been called the 'activists' and the
'muddlers-through' (Kellogg, 1987). The former call for early action, arguing
that trace gases could cause deleterious and irreversible atmospheric changes
long before science fully understands them. The latter urge against action,
arguing that without thoroughly understanding the problem it would try to
solve, society might take costly, unnecessary, or ineffective measures. Judg-
ment of whether action would be worthwhile, however, is beyond the com-
petence of science.
The decision to act or not to slow the release of greenhouse gases belongs
to the realm of what Alvin Weinberg called 'Transcience' (Weinberg, 1967).
It demands complex and personal judgments regarding acceptable levels of risk,
the value of present versus future benefits, intergenerational equity, as well as

Climatic Change 13 (1988) 241-265.

9 1988 by Kluwer Academic Publishers.
242 William U. Chandler

the technical and economic potential for controlling greenhouse gases. The
decision thus can be properly made only by society through its political
institutions.
Scientists, nevertheless, are increasingly asked to inform the climate-change
debate about the potential benefits and costs of action and inaction (NCR,
1983; Firor, 1987; Schneider, 1987). In this respect, science can usefully assess
technical and economic measures for their efficacy. Unfortunately, considerably
less effort has been allotted to understanding means of controlling human
sources of greenhouse gases than to their effects (Edmonds and Reilly, 1983;
Edmonds and Reilly, 1985). Nowhere is this more obvious than when climate
modellers merely extrapolate trends of carbon dioxide emissions - especially
if the resulting scenarios are used to conclude that a greenhouse warming is
inevitable and political action against it futile. This practice is reminiscent of the
errors of energy supply planners of the late sixties and early seventies, who
assumed that energy use (the main source of anthropogenic carbon emissions)
and economic output moved in lock-step, and that energy-use was highly price-
inelastic. This somewhat deterministic outlook has of course been discredited by
the price-induced energy-efficiency response of recent years. Since 1970, the
United States and Japan, for example, have reduced the energy intensity of their
economies by 30% (U.S. Department of Energy, 1987).
Surprisingly, the potential role of energy-efficiency in ameliorating climatic
change due to increasing atmospheric carbon has been minimized in the litera-
ture. This is particularly true concerning the developing nations, which may
hold legitimate claims to increase absolute levels of energy use. For example,
one influential book concluded '... energy conservation has little meaning for
[developing] countries when per capita energy consumption is, say, only 5% of
that in the United States' (Kellogg and Schware, 1981). Energy efficiency, how-
ever, is not defined by per capita rates of consumption but by the level of ameni-
ty supplied per energy unit. Energy is a means, not an end, and as it turns out,
developing nations are often exceedingly inefficient in energy use. Thus, it is
possible that economic growth could occur with considerably less generation of
carbon emissions than the trend-extrapolation scenarios suggest. Indeed,
economic growth will likely be retarded by energy-use inefficiency. But it
remains true that economic growth could be sufficient to cause a major atmo-
spheric warming. To assess the overall potential for policies that affect energy
use to reduce the growth of carbon emissions and their impact on incomes, one
needs a consistent, transparent, theoretically sound, and reproducible frame-
work, or model. Any model - and an extrapolation or hunch in one's head is a
model - which does not meet these criteria not only lacks rigor but is invalid.
This paper asks whether it is technically and economically feasible for a
major developing country, China, to reduce significantly its current and
projected energy-intensity and its expected carbon emissions. China is a logical
choice for such a study: It represents 20% of the human population; it possesses
 Assessing Carbon Emission Control &rategies: The Case of China 243

and uses large quantities of coal; and it will likely be adversely affected over
large areas of its interior by a 'drying-out' if global atmospheric warming occurs
(Chino, 1985; Zong-ci and Kellogg, in press).
This paper briefly describes and applies the well-known Edmonds-Reilly
energy-economic model to assess (1) the relative effectiveness of Chinese initia-
tives to reduce the global concentration of atmospheric carbon dioxide to the
year 2075, (2) the effectiveness of similar international efforts, and (3) the effect
of each of these initiatives on Chinese income levels. The scenarios have been
run to the years 2000, 2025, 2050, and 2075, with results presented for the last
year. Measures considered include family planning, fossil fuel taxes, mandatory
or technical energy efficiency improvements, and a combination of these
measures.
The remaining measures provide a context for evaluating the relative con-
tribution of and costs to the Chinese people of carbon emissions control. Four of
these are energy efficiency scenarios - one for western developed nations, one for
developing nations including China, and two for the entire world. The most
important of these asks, What if the countries of the world organized to reach
by 2025 current levels of energy efficiency in Japan or Austria, and then
reached the 2025 Japanese (or Austrian) level by 2075?
Others include a world 'low-cost solar energy' projection, a ban on coal use
and export after the year 2000 by the Soviet Union, the United States, and
China, and a ban on coal use after 2000 by the entire world. In addition, tests of
the sensitivity of the model to energy prices and changes in critical economic
variables have been conducted and are briefly discussed.
The paper goes beyond traditional macroeconomic analysis of future energy
demand, however, to consider microeconomic-engineering analysis, a major
energy-policy tool which emerged in the early seventies. This brief application
tests the realism of lower energy demand scenarios by comparing extant energy-
efficiency opportunities with those implied by macroeconomic scenario analysis.

2. The Chinese Fossil-Energy Context

The difficult issues of equity and effectiveness in carbon dioxide control

converge around China because it remains a developing country, faces rapid
population growth, and produces an unusually high rate of carbon emissions per
unit of economic output. Despite annual economic growth of almost 10% in the
post-Mao years, Chinese per capita income in the late eighties averaged only
about $600, 4% of the U.S. level (World Bank, 1985). Chinese leaders - and
others - would surely question the equity of measures that benefitted the global
environment at the expense of economic growth in their poor nation.
Similarly, Chinese per capita energy use is only about one-third of the world
average, and less than 7% of the U.S. rate. (World Resources Institute, 1987; see
Table I). Though China's population represents one-fifth of the world total, its
244 William U. Chandler

TABLEI: Chinesepopulation,income,and energyuse in perspective.

Country/world GNP Energy(GJ) Population

US$ per Growth Per cap. Per 1000 (rail.) Growth

cap. (%/yr) '755GNP (%/yr)

(1984) (1973-84) (1984) (1987) (1985-90)

China 310 6.6 19 69 1085 1.18
Brazil 1710 4.0 51 16 179 2.07
Japan 10,390 4.3 111 18 130 0.51
South Korea 2090 7.2 44 45 42 1.89
United States 15,490 2.4 280 32 242 0.86
WORLD 24001 na 55 na 5000 1.63

1 Estimated.
Source: International Institute for Environment and Development and World Resources Institute,
1987.

fossil-energy carbon emissions amount to only one-tenth of the global total.

Measures to curb fossil-energy use based on per capita rates of consumption
would probably not be considered equitable.
Advocates of controlling carbon emissions might nevertheless call for sub-
stantial efforts in China, arguing that its population growth is large and its
economy is energy-intensive. China over the next two decades may add almost
as many people as the entire current population of the United States (Year 2000
Group, 1984). That rate could be higher if China abandoned its ambitious
family planning efforts. At the same time, only Hungary and the oil-rich Mid-
Eastern nations use energy more inefficiently than China. In 1984, China used
over twice as much energy per unit of economic output as the United States and
four times as much as Japan. (IIED, 1987)
China has made major energy-efficiency gains recently as energy use per unit
of economic output declined annually by 3.4% between 1979 and 1985. Interest-
ingly, the income elasticity of demand for energy in China has averaged 0.5
during the eighties, about half the level that conventional wisdom would predict
for a nation at China's stage of development (Lu, 1986b; Institute of Energy
Economics, 1987; Xu, 1988). Coal use in China, nevertheless, has grown at an
annual rate of 8% per year since the 19 50s, and coal now supplies three-quarters
of China's energy (Lu, 1986a; Bentjerodt, 1985). China thus has the highest rate
of carbon emission per unit of economic output in the world, as well as a rapidly
growing absolute rate of carbon emissions (see Figures 1 and 2).

3. The Edmonds-Reilly Long-Term Energy-Economic Carbon Dioxide Model

The Edmonds-Reilly long-term, energy-economic model is applied here to

assess the effectiveness and implications of policies for controlling carbon
dioxide emissions. It is a partial-equilibrium model, meaning that it solves for
 Assessing Carbon Emission Control Strategies." The Case of China 245

1.0

0.9

0,8
II.
z 0,7

o 0.6
0
O

-~ 0.5
o=
-E
8 0.4

0.3

0.2

0.1

0
France Japan FRG Italy UK USA USSR Poland DRG China

Fig. 1. Carbon emissions per dollar of output, selected countries, 1983.

Source." author.

1.3
1.2
1.1
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0,3
0.2
0.1
0 ~
1950 19~55 19~60 19~65 19~70 1975 19~80 1985

SOURCE: R. Ratty, UN data.

Fig. 2. Carbon emissions 1950-84.
Source. R. M. Ratty, personal communication.

equilibrium between supply and d e m a n d and between the energy sector and the
rest o f the economy. The energy sector is represented in considerable detail,
though the overall e c o n o m y is not.
246 William U. Chandler

The model's wide and continued use stems from three attributes, which are
prerequisites for valid economic modelling. First, it is based on sound economic
principles. Economic activity, or GNP, drives energy demand. GNP, prices,
energy tariffs and trade barriers, and technological change also affect energy
demand and, therefore, anthropogenic carbon emissions (Chandler, 1985). The
world economy is simply represented, as is appropriate for modelling over the
very long term. Exogenous variables such as population growth are difficult to
predict accurately, and assumptions for price and income elasticity of energy
demand are controversial (Gibbons and Chandler, 1981).
Second, the model is transparent. The key assumptions and economic
principles that drive the model have been clearly articulated. The major
assumptions are open to view and modification. The model's authors con-
structed a tool that would permit the user to explore a range of alternative
scenarios using an array of policy options.
Third, the model is reproducible. The modelling philosophy, principles, and
technical detail have been well-documented. It is also readily obtainable,
permitting a variety of users to conduct policy experiments in a common frame-
work.
The uncertainty inherent in making long-term projections led Edmonds and
Reilly to base their model on only the most basic behavioral principles. Thus,
they took a 'minimum' approach to modelling, specifying six minimum require-
ments for the model:
(1) Disaggregation by fuel type;
(2) long-term applicability;
(3) global scale;
(4) regional detail;
(5) supply-demand equilibrium;
(6) carbon dioxide flow accounting (Edmonds and Reilly, 1983).
Separate modules, or subroutines, determine initial values for economic growth,
energy demand, energy prices, and energy supply (see Figure 3).

3.1. Energy Demand

Global energy demand in the Edmonds-Reilly model is the sum of energy use in
9 political-economic regions, including one adapted here to represent China.
Economic growth, or GNP, fundamentally drives regional energy demand, and
is itself a dependent variable determined by rate of growth in labor force and
labor productivity. That is,

cNPo= wij . LP , (1)

where W is the number of workers and LP is labor productivity per worker for
 Assessing Carbon Emission Control Strategies: The Case of China 247

•
L _ ~ J .................. ~
REGIONAL
IENERGY I

, l w!

i k_ l M/

~ SUPPLY

Fig. 3. Schematicof the Edmonds-Reillymodel.

Source." Edmondsand Reilly, 1983.

region i in year j. Values for these independent variables are assumed

exogenously and can be varied by the model user.
Energy demand accounting for the countries of the Organisation for
Economic Cooperation and Development (OECD) is organized by energy
service sector: residential/commercial, transportation, and industry. Energy
end-use in all other nations is aggregated in a single sector. In the OECD
buildings and transportation sectors, energy service demand (Es~) is related to
energy price (P), an index of per capita G N P ( X ) , and population (Pop).
Specifically:
Esk = prkPk * X ryk * Pop (2)
where rpk and ryk are price and income elasticities of demand. Energy service
demand in the non-OECD economies and in the OECD industrial sector
demand is formulated slightly differently:
Es k = p;pk . xryk . y , (3)
248 William u. Chandler

with an index for economic activity Y replacing population as a principal deter-

minant of demand.
Total demand for an energy carrier is then calculated from the demand for
service type, the carrier's share of service provided (see 'Supply', below), energy
required per unit of service, and a factor representing technological change in
energy required per unit of service demanded. Technological change is the
exogenously-assumed rate of improvement in energy efficiency that may occur
in addition to price response. It reflects changes that improve energy-use
efficiency beyond the effect of price response. This factor is justified empirically;
such a mechanism appears to have reduced U.S. industrial energy consumption
per unit of output by about 1% per year between World War II and the early
seventies, despite declining real energy prices (OTA, 1981). The value for this
factor is selected by the model user.
Total regional energy demand by service type is the sum of demand by
consuming sector. Total primary energy demand (e.g., coal), is then calculated
from demand for energy services and assumptions regarding the efficiency for
conversion to energy carrier (e.g., coal to electricity).

3.2. Supply

The type of energy service demanded by region and the type of fuel used to
create it have obvious importance for modelling CO2 emissions: Carbon release
per unit of energy produced differs by a factor of two between fuels such as
natural gas and oil shale. The Edmonds-Reilly model accounts for several 'fuel'
types: conventional and unconventional oil, conventional and unconventional
gas, coal, shale oil, biomass, solar electricity, nuclear electricity, and hydro-
electricity. The shares of energy provided by the secondary fuel-types are deter-
mined in this model on the basis of cost and income. That is, fuel types have
both income and price elasticities of demand. Logistic functions (which simulate
market penetration curves) determine the actual shares of services provided by
non-renewable energy sources. This non-economic aspect of the model has been
criticized (NRC, 1983).
Edmonds and Reilly maintain that although this market-share sub-model
does not incorporate economic behavior but simply extrapolates production
rates for energy supply from an assumed resource base, it nevertheless provides
an accurate description of reality, or at least one that can be legitimately used for
modelling purposes. The alternative, they argue, is to use 'discovery process'
modelling which must incorporate not only "prices, resources, and discovery
constraints.., but also.., expected future prices" (Edmonds and Reilly, 1985).
Thus, these more 'intellectually satisfying' models do not provide greater reli-
ability.
The logit market-share model for non-renewable energy production is based
on an assumption for the total quantity of the exploitable resource, parameters
 Assessing Carbon Emission Control Strategies: The Case of China 249

which determine the rate of exploitation, and time. Thus, production rates are
assumed exogenously, an aspect which makes the model both 'non-economic'
and flexible. Although supply behavior is mechanistic, one can alter the
assumptions about resources and supply rates to assess various alternative
futures, in effect shifting the supply curve.
Supplies of renewable resources are treated somewhat similarly. In the logit
formulation, market share is determined by parameters that set a 'breakthrough
price' or point at which the resource becomes competitive, and by the ratio of
the cost of the resource to that of its non-renewable competition.

3.3. Supply-Demand Equilibrium

A simple, iterative process is used to clear markets for oil, gas, and coal, while
prices for non-tradeable sources such as nuclear power are assumed
exogenouslyJ Beginning with an arbitrary set of prices, a 'price preprocessor'
selects optimal energy supply configurations. Next, a procedure tests the differ-
ence between the initial levels of world energy supply and demand. This 'gap' is
measured as:
Xi -- In Q~ - In QSi, (4)
where i represents oil, gas, and coal, and Qo and QS represent quantities of
energy supply and demand. An estimate is then made of the price changes
needed to close the 'gap' by finding the difference between the relevant set of
price elasticities of demand and of supply, or:
dXi = Sumj(U 0 - V~j) dln Pj, (5)
where U and V are the demand and supply price elasticities and j is oil, gas, or
coal. A calculation of the necessary price changes is then made by inverting a
matrix of the demand and price elasticities (and cross elasticities) by fuel type.
Note that the regional elasticities are determined themselves in the iterative
procedure.

3.4. Modelling Issues

The importance that price plays in the model can be seen with a test of the
sensitivity of energy intensity per unit of GNP to price changes. For example,
taxing global fossil fuel use by 200% over market prices reduces the energy
intensity - that is, energy per unit of economic output - of the global economy
in the year 2025 by 25% compared to the Base Case.
Assumptions for the capital, labor, or total costs of energy supplies are open

I Demand for non-tradeable sources such as nuclear, solar, and hydroelectricity does not undergo
the equilibrium-finding process, being assumed to equal supply in the regional demand functions.
250 William U. Chandler

to change by the user. Coefficients that include the capital requirements for
nuclear power plants, for instance, can be increased to reflect recent escalation.
Also, resource availability can be reduced to reflect nuclear moratoria or de
facto moritoria. Nuclear energy resources in this set of scenarios were restricted,
reaching only the equivalent contribution of hydroelectric power worldwide by
the year 2050.
The model does not serve short-term planning needs well. It was not of
course designed for short time horizons, but policy decisions are almost always
made on this basis. The model serves best in this regard to screen for the most
useful 'solutions' to problems as they are defined. Short-term policy making
can then be guided by the more common - and more commonly available -
demand models. For example, Edmonds-Reilly can be used to discover whether
a particular set of policy objectives - reducing emissions while maintaining
incomes, for example - can be better served by energy taxes or population
control measures, if at all. But if one wanted to determine the optimal mix of
specific regulations, one would have to resort to engineering-economic or input-
output models. In any case, the value of extrapolating or assuming end-use
detail at the level, say, of ton-kilometers of rail and truck freight per year would
probably add little reliability to the aggregated or macroeconomic approach to
modelling many decades into the future.
A more troublesome issue is the inclusion of only one end-use sector for non-
OECD nations. This shortcoming is diminished somewhat by the fact that
commercial energy use in Eastern Bloc and developing nations, including China,
is dominated by the industrial sector. Industry uses over 60% of all Chinese
commercial energy. Over the long run, however, industry will decline as a share
of economic activity in these economies. The replacement of non-commercial
with commercial fuels in the buildings sector will also diminish the fuel-share of
industry, a likelihood which points to the weakness of using a single demand
sector for any economy.
One may also question the validity of using the model to analyze behavior in
centrally-planned economies such as China. For example, the price of coal,
which supplies three-quarters of China's commercial energy, is determined in
the model by international competition, while in reality coal in China is
allocated almost without regard to price. (Lu, 1986b; Bentjerodt, 1985) This
issue does not, however, present serious theoretical problems (from a Western,
mainstream economics point of view). Elementary economics holds that an
efficient, centrally-planned economy at best approximates supply and demand
equilibrium in a competitive economy (Samuelson and Nordhaus, 1985). This
means that China, if it were to continue under central planning, and if it were
efficiently managed, would follow approximately the same resource-use patterns
as if it were a market economy.
Energy prices remain heavily subsidized for Chinese consumers, however,
and long-run equilibrium as defined by western economists is a distant vision at
 Assessing Carbon Emission Control Strategies: The Case of China 251

best. But the Edmonds-Reilly model can be adapted to reflect this reality, even
to extrapolate similar conditions, however unsustainable. As noted above, the
model facilitates the application of taxes to energy prices. Negative taxes can be
applied to represent subsidies - direct or through price controls - to energy
consumers. Chinese energy-using behavior can thus be studied via the
mechanism of price response, which is widely accepted as a fundamental
behavioral relation not limited to market-oriented societies. Indeed, there may
be little difference in principle between centrally planned economies and the
regulated utility sectors of the U.S. economy, which deliver half of all U.S.
energy supplies to final demand. Also, as noted earlier, price elasticities of
demand can also be modified to reflect regional differences. There is the further
difficulty of a multi-tiered price system, such as the Chinese currently use,
which permits one price for energy supplied or purchased up to a given level,
and a higher one for above-quota supply or demand. In this case, the prices
specified in the Edmonds-Reilly model can at best approximate an average for
all pricing tiers.
Another issue pertains to the partial nature of this equilibrium model. It does
not incorporate major economic changes in the non-energy sector of the
economy beyond the rudimentary relationships of energy demand to economic
growth, energy prices, and technological change. The model cannot incorporate
changes in savings rates, for example, which might affect capital formation.
Capital formation can have important effects on energy efficiency, for to a large
degree, energy savings are achieved through the substitution of capital for
energy. At the same time, however, capital formation and labor or energy
productivity are analytically separate. Productivity improvements, for energy as
well as for labor, can be capital saving (Samuelson, 1970). Efficient industrial
motors in Brazil, for example, avoid the need for new electric capacity at a
capital cost of about $300 per kilowatt, whereas new hydroelectric supply costs
upwards of $2000 per kilowatt (Geller, 1985). Thus, incorporating capital
formation in a long-term model would simply permit one to assume results with
respect to labor and energy productivity. One can simply inject such effects
directly using the mechanisms in the Edmonds-Reilly model for altering
assumptions of labor and energy productivity.
Still, a short-term model would provide more insight in this regard, and the
importance of these variables should not be minimized. A sensitivity test on the
Base Case which cut labor productivity growth - as a surrogate for capital
formation - to only l% per year worldwide, reduced global GNP in the year
2025 by more than two-thirds. In this case, global carbon emissions rapidly fall
below current levels and stay there. The implication is that if capital formation
(or some aspect of it, such as savings rates) slows, economic growth slows and
growth in carbon emissions would slow (or decline) as well. But an increase in
capital formation could either increase carbon emissions by increasing long-term
economic growth, or reduce them by making the economy more energy-
252 William U. Chandler

efficient. The actual result would depend on changes in technology and

consumer preferences, which the model tests.

4. The Potential Impact of Carbon Dioxide Control Policies in China

Twelve scenarios below compare the effectiveness of carbon dioxide control

policies in China and their impact on Chinese living standards. See Table II for
major assumptions underlying the cases (for further details, see Edmonds and
Reilly, 1985b).
The model is initialized to the year 1975. A Base Case scenario is constructed
from 'Medium' scenario assumptions in a recent U.S. Department of Energy
study (U.S. Department of Energy, 1986). These two 'cases' provide a context
for assessment.
(0) Worm Reference, 1975: the initial calibration period.
(1) Worm Base Case: a scenario generated from 'medium' values for all
assumptions, with these values being selected by Edmonds and Reilly. It
assumes a 1 percent annual rate of energy efficiency improvement in all
regions.

TABLE II: Major assumptions for the scenarios

Population: World total, year 2075 = 8 billion with Chinese Family Planning, 8.4 billion without.
China population = 1.2 billion at equilibrium in the Family Planning scenario, 1.6 billion in all
others.

Labor Productivity: Developed countries=0.017 growth per year; developing countries=0.029

growth per year.

Income elasticity of energy demand." Initial values of 1, 1.25, and 1.4 for developed western,
developed eastern bloc, and developing countries, respectively; converging toward 1 by the year
2075.

Price elasticity of energy demand." -0.7 for all nations for all sectors.

Energy price-GNP Feedback elasticity: -0.15 and -0.2 for developed and developing countries,
respectively.

Fossilfuel resources: Oil = 21,250 exajoules (EJ), Natural gas = 13,950 E J, Coal = 271,000 EJ.

The model is altered to permit assessment of the People's Republic of China

by shifting the relatively small populations of Kampuchea, North Korea, and
Vietnam out of the Centrally Planned Asia region into the South East Asia
region, which includes India, Pakistan, Thailand, and Afghanistan. Four
scenarios deal with China and Chinese policies alone.
(2) China Family Planning: a scenario identical to the Base Case, except that
 Assessing Carbon Emission Control Strategies: The Case of China 253

China's population is assumed to stabilize at 1.2 billion rather than at

about 1.6 billion as in the Base Case. 2
(3) China Revenue Shift. differs from the Base Case by assuming that a tax is
imposed to triple the price of coal, oil, and natural gas in China in order
to discourage fossil fuel use. Revenues are assumed to be 'recycled' to the
economy or substituted for other revenue sources.
(4) China Energy-Efficiency Standards: incorporates a 2% annual rate of
improvement in Chinese energy efficiency, 1% elsewhere. This scenario
assumes that energy efficiency improvements are effected by fiat or by tech-
nological change, that is, either by regulatory mechanisms or by evolution of
energy-use technology.
(5) China Combination:a scenario combining non-economic energy efficien-
cy improvements of 2% per year, low population growth (1.2 billion
equilibrium level), and price-induced energy efficiency improvements stemming
from a tax raising fossil fuel prices in China to 300% of 1975 levels.
The remaining seven scenarios evaluate multi-lateral or non-Chinese carbon-
control initiatives:
(6) Worm Low-Cost Solar Energy." tests a case in which solar energy would,
by 2025, cost only $7.43 per gigajoule (1975 $), half the assumed cost in the Base
Case for the year 2050.
(7) OECD Energy-Efficiency Standards." assumes the OECD nations attain
2% per year energy productivity improvements (with 1% per year elsewhere).
This scenario helps differentiate the contributions to ameliorating the CO2
problem that Western countries can make relative to developing and centrally
planned nations.
(8) U.S., Soviet Union, and China Coal Ban: tests the ability of these three
major coal consumers to address the carbon dioxide problem with a three-nation
ban on coal use and exports after the year 2000.
(9) Developing Nations Efficiency Standards: a scenario of 2% per year
energy productivity growth in the developing world, with a 1% rate in the
OECD and Soviet Bloc countries.
(10) Worm Revenue Shift: a case that raises world energy prices to 300% of
1975 levels to promote energy efficiency.
(11) Worm Coal Ban: a test of the impact of banning coal use worldwide after
the year 2000.
(12) Worm Efficiency Standards: a case in which world-wide energy efficien-
cy improvements - through technical change or by agreement - averaged 2%
annually, permitting the world to reach current Japanese or Danish levels of
energy efficiency by year 2025.

2 Note that 1.2 billion is the current official goal of the Chinese government, despite evidence that
enforcement of the Chinese family planning policy appears now not to be stringent.
254 William U. Chandler

4.1. Results

Global fossil-fuel carbon emissions in the Base Case scenario would by 2075
reach over 19.1 gigatonnes per year (a 1.5% annual rate of increase), almost four
times the level of the 1980s. If 55% of the carbon dioxide remained in the
atmosphere, the atmospheric concentration in the year 2075 would be about
600 ppm, compared with 345 ppm currently. 3 Global per capita income would -
in the absence of climatic feedback - reach almost twice current U.S. levels.
Chinese incomes would average $7200 (1975 $), or about the 1987 U.S. average.
The China Family Planning scenario would reduce global carbon emissions
by 800 megatonnes in the year 2075 compared to the Base Case rate. The global
atmospheric carbon dioxide concentration in this case would be reduced slight-
ly, to 594 ppm in the year 2075. Chinese per capita income in that year would
increase relative to the Base Case by more than $200 (see Table III). More
importantly, year 2025 Chinese per capita income would have climbed $300
higher than in the Base Case. (Note that the scenarios are not cumulative in
assumption or result.)
Tripling the price of fossil fuels in China by taxing coal, oil, and natural gas
would reduce year 2075 carbon emissions by 8%, or 1.6 gigatonnes, again,
relative to the Base Case. If it is assumed that the tax revenues from energy are
simply removed from the Chinese economy, per capita annual Chinese income
is reduced compared with the Base by about $ 3500. This effect results from the
"Energy Price-GNP Feedback Elasticity', which in the Base Case is set for China
at -0.2. This elasticity may be inappropriate for price increases due to govern-
ment taxation: The money can be rebated or used to substitute for other
revenues. However, there is little empirical evidence of an appropriate value for
this elasticity. Reducing it in proportion to the tax, which increases energy
prices by a factor of three, would yield an income level slightly higher than the
Base Case. This change presumably results from downward pressure on energy
supply prices, which increases GNP through the same feedback elasticity.
A scenario of mandated energy efficiency - alternatively, a 'technological fix'
through innovation and evolution - in China would reduce global carbon emis-
sions somewhat more than tripling the price of fossil fuels in China. Chinese
incomes increase $100 above the Base Case, less than in the China Revenue

3 Carbon dioxide emissions are translated into accumulated atmospheric concentrations applying
the integral
At = Ao + 0.471 C ( e rt - l)f/r

where A is the atmospheric concentration in the year t projected, A o is the concentration in the base
year (1975 = 339 ppm), C is the annual rate of emissions in tonnes in the base year (1975 = 4.7
billion tonnes), r is the rate of growth in emissions calculated from the initial and end-point emis-
sions rates, and f is the fraction of emissions remaining airborne. Various analyses suggest that a
reasonable estimate o f f is 0.55, though the range may be considerably wider, and, more important-
ly, may increase with time if the ocean sink for CO2 becomes saturated (Edmonds and Reilly, 1985).
 Assessing Carbon Emission Control Strategies: The Case of China 255

Table III: Year 2075 scenario results: Chinese and international carbon dioxide control policy
options

Scenario Global carbon Chinese income Atmospheric

emissions (C in per capita concentration 1
gigatonnes) ($1975) (ppm)

(0) 1975 actual 4.5 360 339

(1) World base case 19.1 7236 601
(2) China family planning 18.3 7447 594
(3) China revenue shift 17.5 7413 587
(4) China efficiency standards 17.1 7336 583
(5) China combination 16.8 6087 581
(6) World low-cost solar energy 16.3 7398 576
(7) OECD efficiency standards 16.4 7428 577
(8) US-Soviet-China coal ban 9.5 6277 512
(9) Developing nation eft. stds. 10.8 7561 525
(10) World revenue shift 9.1 7069 508
(I 1) World coal ban (after 2000) 5.3 5450 466
(12) World efficiency standards 5.9 7739 473

1 Assumes an atmospheric carbon retention rate of 0.55.

See text for definitions.
Source: author.

Shift scenario. The China Combination scenario produces slightly results - an

additional 300 megatonne carbon reduction with little income effect, as long as
revenues are recycled.
The international scenarios were, not surprisingly, considerably more effec-
tive in reducing future carbon emissions. A scenario of high efficiency for
developed Western nations acting alone, the OECD Efficiency Standards
scenario, reduces carbon emissions by 2.7 gigatonnes compared to the Base
Case, and holds the atmospheric concentration in the year 2075 to about 575
ppm. A much greater reduction - more than 8 gigatonnes (to 525 ppm) - results
in the Developing Nation Efficiency Standards scenario. This case assumes a 1%
energy efficiency improvement rate in the Western nations and a 2% rate in the
rapidly growing economies of developing nations. Chinese income levels would
rise slightly in either scenario.
The combined efforts of the United States, the Soviet Union, and China to
ban coal use both for domestic use and export after the year 2000 would reduce
year 2075 Base Case carbon emissions by 9.6 gigatonnes. Atmospheric carbon
dioxide concentration would thus be held to just over 510 ppm for the year
2075. Chinese incomes would fall almost $1000 compared to the Base Case, and
thus China would not reach current U.S. income levels by the year 2075.
A global ban on coal might yield an atmospheric CO2 concentration of about
465 ppm of carbon dioxide. In that case, Chinese might lose almost $1800 in per
capita GNP, compared to the Base Case scenario.
256 William U. Chandler

An early, technological breakthrough that reduced the cost of solar (or

nuclear) energy to low levels would not appear to keep carbon dioxide concen-
trations much below 600 ppm. The Worm Low-Cost Solar scenario assumes
that the cost of solar energy is somehow reduced steadily until the year 2000
when it reaches the equivalent of $7.45 per gigajoule ($1975 dollars). This price
level was chosen arbitrarily because it represents a halving of the price level
expected by Edmonds-Reilly. Similarly, it was assumed that the date by which
the lower price would be achieved was advanced by 25 yr. Though this scenario
appears optimistic, it suggests that because new energy sources suffer a time lag
in market penetration, they may be considerably less effective in reducing
carbon emissions than available energy efficiency options. These conclusions
could just as well apply to efforts to promote nuclear power (see Figure 4 for
selected scenario results).
Non-fossil fuel energy sources are often touted as solutions to the carbon
dioxide greenhouse problem. Advocates of hydroelectric power, nuclear power,
and solar energy sometimes argue that the climatic benefits of using these energy
systems outweigh any other negative aspects. But a scenario (not presented here)
in which China completely abandoned new sources of hydroelectric power
would have little effect on the greenhouse problem. The difference amounts to
100 megatonnes of carbon, or 1 ppm atmospheric carbon dioxide concentration
by the year 2075.
An important point must be noted regarding the sensitivity of the model to

26

24

22

20

18
>-
16

14

12

10

0 i i i i i i
1980 20'00 20'20 20'40 20'60 20~80 21O0
(Year)
1. World Bose Case ......... 7. 0ECD Efficiency ..... 6. World Solar Case
--- 11. World Coal Ban 12. World Efficiency

Fig. 4. Global carbon emissionsby scenario,

Source: author.
 Assessing Carbon Emission Control Strategies: The Case of China 2 57

physical and political and economic assumptions. If one assumed a 50% rate of
atmospheric carbon retention instead of the 55% rate used above, the results for
the Base Case for example, would be a carbon dioxide concentration of 577 ppm
in the year 2075 instead of 601. Though it does not change the relative value of
policies, the atmospheric retention assumption can overwhelm the effect of
policies affecting China alone.
These results, predictably perhaps, suggest that no nation alone, not even
China, can decisively affect the global CO2 problem. From a policy perspective,
this finding is reason for concern; obtaining a concensus for effective and timely
action on so profound and controversial an issue will be diffficult. But more
important is the result suggested that only energy efficiency scenarios both
significantly reduce carbon emissions and increase Chinese per capita incomes.
That is, climatic change could be slowed by enacting measures that are
apparently in China's own self-interest. The same conclusion applies to the rest
of the world.

5. Discussion: Projections and the Two Paradigms

Scenarios of future energy demand always convey a certain sense of unreality.

This is particularly true when they draw apparently inconsistent conclusions.
That is the case in the scenarios above which suggest that 8 billion people can
achieve per capita incomes equal at a minimum to current American levels,
despite demands on the natural environment which may be unsustainable. Even
if atmospheric carbon dioxide levels double, the pre-industrial levels do not
disrupt the global economy through the effects of climatic change, it is difficult
to envision physically how four times the current quantity of energy resources
could be supplied, much less sustained, without other major environmental
disruptions which could in turn disrupt regional economies. One can always, of
course, appeal to technological breakthrough to provide a safe, clean, and
relatively inexpensive way of supplying so much energy. However, one could
just as easily assume economic stagnation or collapse.
It is worth reiterating that these scenarios are not predictions but are policy
experiments. What is most important about them is certainly not their precision,
but the direction and magnitude of the impacts they suggest.
A growing literature turns away from the economic or behavioral paradigm
presented here to construct energy end-use technology scenarios showing how
human needs could be satisfied at far lower energy demand levels. This
approach assesses the future from a microeconomic and engineering perspective,
as opposed to the macroeconomic approach of Edmonds-Reilly. To a certain
extent, the two paradigms depict possible worlds (microeconomic-engineering)
and probable worlds (macroeconomic).
Rather than competing as world views, however, the approaches can serve to
check or even complement each other. While it is implausible to make 100-yr
258 William U. Chandler

forecasts of technical details such as the number ofton-km of freight transported

or the number of passenger-km of air transport, it is instructive to estimate how
a larger human population could more sustainably and equitably avail itself of
the amenities of energy supply. A recent study from the World Resources
Institute (WRI) makes such calculations, and the results have relevance for
policy-makers who would use energy efficiency as a tool in nations like China to
reduce carbon emissions. Their technical and economic conclusions have been
reached by many others (Gibbons and Chandler, 1981; Ross and Williams,
1981; SERI, 1981; Kelly, 1986).
The WRI study avoids almost entirely the question of how human behavior
can be changed to bring about energy efficiency. Instead, it calculates for the
year 2030 the energy requirements of supplying persons of developing nations
with what its authors consider 'basic human needs'. For the developing world,
this criterion translates into multiples of recent energy-service usage levels, for
example, 5 times recent levels of hot water use, 20 times as many cars per
capita, 20 times as much air travel, 15 times as much steel, 8 times as much
fertilizer, and so on (Goldemberg, et al., 1987). The study calculates energy
requirements based on currently-available technical means that save energy for
less cost than the current marginal prices of fuel supplies. The result cor-
responds to an annual energy efficiency improvement of about 2.5% per year
over the period. Energy efficiency would improve much faster in developed
nations, however, where it would average 3.4% per year.
The microeconomic-engineering approach does not pretend to make predic-
tions about future energy-using behavior or how it can be modified. But it
provides a basis for assessing the technical plausibility of achieving high levels of
energy efficiency improvement over long periods - the kinds of improvements
achieved in the macroeconomic energy efficiency scenarios described above.
Comparing the known rates of energy use in specific end-uses in China with the
best available, economically justifiable technologies, it is evident that China
could maintain a 2% rate of energy efficiency improvement for almost eight
decades - without any new technological breakthroughs. This calculation can be
reproduced more simply: China uses four times as much energy per dollar of
economic output as Germany, France, Italy, or Japan (Bentjerodt, 1985;
Chandler, 1985). Simply evolving toward an economy as efficient as Japan's or
Austria's today would permit China to maintain the 2% rate of energy efficiency
improvement for 80 yr (see Figure 5). An even higher rate of efficiency improve-
ment could be sustained by applying the best technology currently economically
justified.
These facts give rise to a question. Why should China's economy, almost all
of which will be built or replaced in the coming decades, take almost a century
simply to reach Japanese or Austrian energy-efficiency levels of the 1980s? That
is exactly the prospect painted by the Base Case scenario. It is sometimes argued
that developing nations like China have higher energy-intensities because of
 Assessing Carbon Emission Control Strategies. The Case of China 259

70

5O

v
4o
O
g
'~' 30
O.

20 Japan in 1985 "'"-...

10
Scenario 4

0 I I I
1980 1990 2000 20=10 2020 2030 20=4o 20=50 2060 2070 2080
Fig. 5. Projected energy intensity of the Chinese economy.
Source. author.

their stage of development. That is, most of their G N P comes from industry
rather than the more labor-intensive services sector, and therefore one would
expect China to be energy-intensive compared to developed countries. But, as a
careful World Bank analysis has shown, China's industrial sector is very energy-
intensive compared to developing nations. The share of GNP from industry (as
well as the mix of industry) in South Korea is similar to that in China, for
example, but the latter uses twice as much energy per unit of output as the
former (Bentjerodt, 1985).
It would seem that the Base Case scenario and all simple extrapolations
assume that the Chinese will make some very unintelligent decisions. For
example, it implicitly assumes the Chinese continue to build very small and
inefficient metallurgical facilities, which is a major reason why their industrial
sector is inefficient. The world steel industry, which consumes about 6% of all
commercial energy used in the world, requires almost 16% of all commercial
energy used in China. Yet, no nation other than India uses more energy per
tonne to make steel than China (see Figure 6). The Chinese rate is 50% higher
than in the United States, 60% higher than in Brazil, and twice that in Japan
(Institute for Energy Economics, 1987). The differences are easily attributed to
technology. The Chinese not only make steel in a number of small facilities, but
they use the ineffmient open-hearth furnace for a third of their production, while
Japan has phased out that technology. It is unlikely that an expansion of the
Chinese steel industry - China in recent years has had to import one-eighth to
one-fourth of its steel - would be based on small-scale plants or open-hearth
technology.
260 William U. Chandler

The Chinese chemical and building-materials industries also compare poorly.

Synthetic ammonia production requires over twice as much energy per tonne as
new plants built elsewhere (Bentjerodt, 1985). Similarly, most bricks are
produced in small-scale rural operations that use four times as much energy as
new kilns built elsewhere. Again, there is no logical reason to assume that the
shift from such inefficient facilities will be gradual during a time when the
Chinese economy is doubling every 7-10 yr. Indeed, unless the Chinese make
irrational decisions, the Chinese economy could experience a sharp discontinui-
ty in the energy-intensity of their industry as new facilities replace or outnumber
old, inefficient ones. In other words, one should not expect both rapid economic
growth and continued high energy-inefficiency.
Similarly, the Chinese transportation sector is extremely inefficient in energy
use. Chinese trucks use about twice as much fuel per tonne-kilometer as Ameri-
can trucks of comparable payload (see Table IV). This inefficiency is com-
pounded by load-factors that average about half that of trucks on U.S. interstate
highways (Bentjerodt, 1985). Most trucks are far smaller than the optimum, and
must operate on unpaved roads (one-fifth are paved). Reducing these inefficien-
cies will require a combination of institutional reform, technology transfer, and
capital investment, but improvements to energy-efficiency could be substantial.
Buildings sector energy use is unusually high in China compared to most
developing countries because of large energy requirements for residential space
and water heating. Household fuel is often consumed in small, inefficient stoves.
As the nation develops, its high population density might permit investment in

4O

35

30

~ 2s

0 ! I i I I I I l I I I I I I I I I I
Italy BAT-V FRC Polnd Brazl Franc USSR China
BAT-R Spain Japan Belgm UK USA Czech Austr India

Fig. 6. Energy intensity of steel manufacturing, selected countries, 1980.

Source: author.
 Assessing Carbon Emission Control Strategies': The Case of China 261

TABLEIV: Energyend-use in China: a comparison

End-use Energy-intensity Compared

in China with
(percent)
Steel production -200 Japan
Synthetic ammonia production -250 Best technology
Truck transport >200 U.S.
Cooking -250 U.S. electric
Source: Bentjerodt, China, The Energy Sector.

highly-efficient district heating and cogeneration systems, which are rarely used
today. On the other hand, large quantities of non-commercial energy will
undoubtedly be replaced with commercial energy carriers, and this will increase
m e a s u r e d energy use-levels. The inefficiency of the end-use of non-commercial
fuels should permit consumption rates far below the current total energy levels.
A question of great importance in the buildings sector is the energy-efficiency
of new Chinese appliances. The manufacture of appliances, for example, was
reported to have doubled in 1985 (CIA, 1986). Unless adjusted for technical
change, the Base Case scenario implicitly assumes - via income elasticities of
demand - that Chinese refrigerators will be built like U.S. refrigerators in the
sixties and seventies, when they required over 1700 kWhr per year. But the
average new U.S. refrigerator uses only 1100 kWhr, while the best model
uses less than 750 kWhr per year. The level of efficiency that the Chinese
produce will depend mainly on policy, for virtually no loss of amenity and little
increase in purchase price is experienced in moving to the most efficient models.
The opportunities for energy efficiency are by no means confined to China.
The steet industry worldwide uses an average of 26 gigajoules (GJ) per tonne of
steel produced, far above the 19 GJ used by best-available virgin-ore (BAT-V)
technology (Chandler, 1985). Recycling, moreover, can cut this requirement to
10 GJ per tonne (BAT-R), but 100% recycling is not possible because of impuri-
ties. The thermodynamic minimum for reducing iron from ore is about 6 GJ per
tonne. A two percent annual rate of improvement in the energy efficiency of
steel-making worldwide would necessitate going beyond the thermodynamic
limit by the year 2075, which is of course impossible. The two % rate could,
however, readily be maintained for the next four decades. Further improve-
ments after about the year 2025 would require knowledge that we do not possess
or some steel-substitute that was much less energy intensive.
A 2% rate of improvement for automobile fuel economy would increase the
current world average of 21-23 miles per gallon to about 40 miles per gallon in
the year 2025, and to about 110 mpg in the year 2075. A level of 70-90 miles
per gallon is technically and economically feasible.
An obvious strategy to control carbon dioxide emissions would be to wring
first the maximum cost-effective levels of energy efficiency out of the world's
262 William U. Chandler

economies and then turn to non-fossil renewable energy sources. This strategy is
feasible, but it would be difficult to achieve without artificially restricting fossil
fuel use to force market penetration of new supply technologies. Fossil fuels are
likely to remain competitive for decades with renewable technologies, because
the latter require large capital investments that will probably make them
expensive even compared to scarce fossil fuel. Imposing the 200% tax on fossil
fuel - that is, tripling 1975 prices - might accomplish much of the task. But
effecting such a policy would require policy makers at a minimum to expand
substantially their understanding and acceptance of human-induced climatic
change as a major problem to be addressed.
Another strategy would be to pursue maximum cost-effective energy conser-
vation investments through the year 2025 while simultaneously investing in
basic research in the hope of finding brand new processes that would permit
either energy efficiency improvements to continue at high rates, or low-cost,
benign energy supplies to substitute for fossil fuels. The rationale for this
approach is that fundamental breakthroughs in processes will be necessary, but
that there are no reasons a priori to assume that new energy supply processes
will provide a better solution than new processes that improve efficiency.
There remains the question of how to change energy-using behavior in China
and elsewhere to economically-justifiable levels. The aforementioned WRI study
implies in an appendix that simply maintaining a 1 percent rate of technological
efficiency improvement on top of 'expected' price-induced changes in energy
behavior could effect their scenario for the year 2030. This result can be
approximated with the Edmonds-Reilly models, using price increases for fossil
fuels of 200% (i.e., to 300% of 1975 price levels) while maintaining an annual
1% non-economic, or technological, rate of energy efficiency improvement.
A problem with these assumptions is that energy costs may or may not
increase by so much. And even if they did so, there is no guarantee that
consumer prices will reflect the real cost of energy. Half the world's steel and
coal, for example, are produced in non-market economies where prices rarely
reflect opportunity costs. These nations are invariably the least energy-efficient.
Of course, centrally planned nations could always mandate energy efficiency
improvements, but there is little empirical evidence that this would work. It is
thus not clear how efficiency can be achieved in the absence of signals that
convey to the consumer the economic value of energy (Chandler, 1986). It
should be noted, however, that the Chinese themselves plan to accomplish a
2.2% annual rate of energy-efficiency improvement for the rest of this century,
and have achieved a 3.4% rate during the eighties (Institute for Energy
Economics, 1987; Lu, 1986; Xu, 1988).
Another option would be to sacrifice economic growth or the availability of
some energy services. With the exception of the taxation scenarios, the above
analysis has explicitly assumed that only cost-effective energy-efficiency
measures be taken, ones that would not require any sacrifice of energy services
 Assessing Carbon Emission Control Strategies: The Case of China 263

or any economic loss. But it should be noted that the wealth generated in these
scenarios is enormous: Americans in the year 2075 would be on average seven
times richer than today. The Chinese would be as rich as Americans currently
are. It is conceivable that many citizens - perhaps most - would be willing to
forego some quantity of wealth in order to minimize climatic change. A scenario
in which people in developed countries grew only 3 times richer than they
currently are and maintained a 2% annual rate of energy efficiency improvement
would put year 2075 atmospheric concentration at just under 500 ppm (with all
other assumptions the same as in the Base Case). Their motivation could be
altruism, aesthetic pleasure, or self interest, for many persons undoubtedly will
lose individually if weather patterns shift dramatically. In any case, the choice -
by U.S. standards at least - rests with the populace and is not one that scientists
can or should make for society.

6. Conclusion

The Edmonds-Reilly Energy-Economic model is a simple but useful tool for

assessing the importance of policy responses to the carbon dioxide greenhouse
problem. It provides results for both the effectiveness of measures to control
carbon emissions and the effect of these measures on per capita income. It does
so in a transparent and reproducible framework that cannot be simulated with
rigor or validity outside of formal modelling. The model does have shortcomings,
however, the most important of which is lack of energy end-use detail. There-
fore, any use of the model is best supplemented by engineering-microeconomic
analysis. Combining the two 'paradigms' of micro- and macroeconomic analysis
permits rigor in policy experimentation that otherwise would not be possible.
This application of the Edmonds-Reilly model suggests that China acting
alone could make important but not decisive reductions in global carbon emis-
sions. It also suggests, however, that a truly international effort will be required
to keep atmospheric carbon dioxide levels well below a doubling relative to pre-
industrial levels. From a policy perspective, this finding is reason for concern;
obtaining a concensus for effective and timely action on so profound and
controversial an issue will be difficult.
Most significantly, the results suggest that climatic change can be slowed with
measures that are, for reasons unrelated to climate, in China's own self-interest.
Energy efficiency scenarios both significantly reduce carbon emissions and
increase Chinese per capita incomes. International cooperation for transfer of
technology and knowledge for energy efficiency could provide global benefits.
The choice to ameliorate climatic change, therefore, need not be laden with
unacceptable risks or costs. The decision, however, belongs to citizens and their
leaders. Science cannot and should not make the determination whether carbon
dioxide control policies are worthwhile. At best, science can advise society on
the benefits and costs of such efforts.
264 William U. Chandler

References

Bentjerodt, R., Cao, Weigong, E1-Mekkawy, Abdel, Taylor, Robert, and Fallen-Bailey, Darrel:
1985, China: The Energy Sector, Washington: The World Bank.
Central Intelligence Agency: 'China: Economic Performance in 1985', Testimony before the
Subcommittee on Economic Resources, Competitiveness, and Security Economics, Joint
Economic Committee, U.S. Congress, 17 March 1986.
Chandler, W.U.: 1985, Energy Productivity: Key to Environmental Protection and Economic
Progress, Washington, D.C.: Worldwatch Institute.
Chandler, W.U.: 1986, The Changing Role of the Market in National Economies, Washington:
Worldwatch Institute.
Chino, J.J.: 1985, 'A Methodology for Comparing Energy Distribution Alternatives for the
Technical Development of the People's Republic of China', Ph.D. Dissertation, School of
Engineering and Applied Science, George Washington University, Washington, D.C.
Edmonds, J. and Reilly, J.: 1983, 'A Long-Term Global Energy-Economic Model of Carbon Dioxide
Release from Fossil Fuel Use', Energy Economics.
Edmonds, J.A. and Reilly, J.M.: 1985a, 'Future Global Energy and Carbon Dioxide Emissions',
in John R. Trabalka (ed.), Atmospheric Carbon Dioxide and the Global Carbon Cycle, Spring-
field, VA: National Technical Information Service, Edmonds and Reilly, 1985.
Edmonds, J. and Reilly, J.: 1985b, Global Energy, Cambridge: Cambridge University Press,
Edmonds and Reilly.
Firor, J.: 1987, Director, Advanced Study Program, National Center for Atmospheric Research,
Testimony, Committee on Energy and Natural Resources, U.S. Senate, 9 November 1987.
Geller, H. H.: 1985, 'The Potentiai for Electricity Conservation in Brazil', American Council for an
Energy-Efficient Economy, Washington, D.C.
Gibbons, J.H. and Chandler, W.U.: 1981, Energy: The Conservation Revolution, New York:
Plenum Press.
Goldemberg, Jose, Johansson, Thomas B., Reddy, Amulya K.N., and Williams, Robert H.: 1987
(September), 'Energy for Development', World Resources Institute.
Hafemeister, D., Kelly, H., and Levi, B. (eds.) 1986: Energy Sources: Conservation and Renewables,
Washington, D.C.: American Institute of Physics.
The Institute of Energy Economics (Japan): 1987, 'China's Energy Situation, Present and Future',
NIRA Report, Vol. 4, No. 1.
International Institute for Environment and Development (IIED) and the World Resources Report,
World Resources 1987, New York: Basic Books, 1987.
Kellogg, W.W.: 1987, 'Mankind's Impact on Climate: The Evolution of an Awareness', Climatic
Change 10, 113-136.
Kellogg, W.W. and Schware, R.: 1981, Climate Change and Society." Consequences of Increasing
Atmospheric Carbon Dioxide, Boulder, CO: Westview Press.
Lu Yingzhong: 1986a, 'Long-Term Energy Forecasting of PRC and Strategies in the Development
of Nuclear Energy', Working Paper, Institute for Energy Analysis, Oak Ridge Associated
Universities.
Lu Yingzhong: 1986b, 'Economic Growth and Energy Use in PRC', Working Paper, Institute for
Energy Analysis, Oak Ridge Associated Universities.
National Research Council, Changing Climate: Report of the Carbon Dioxide Assessment Commit-
tee, Washington: National Academy Press, 1983.
Ross, M. and Williams, R.H.: 1981, Our Energy: Regaining Control, Chicago: McGraw-Hill and
Company.
Samuelson, P. and Nordhaus, W: 1985, Economics, Chicago: McGraw-Hill and Company.
Samuelson, P.: 1970, Economics, Chicago: McGraw-Hill and Company.
Solar Energy Research Institute (SER1), Building a Sustainable Energy Society, Washington, D.C.:
U.S. Government Printing Office, 1981.
Schneider, Stephen H.: 1987 (September), 'Insult and Injury', Book Review, Nature 329, 211-212.
U.S. Congress, Office of Technology Assessment (OTA), Industrial Energy Use, Washington, U.S.
Government Printing Office, 1981.
 Assessing Carbon Emission Control Strategies." The Case of China 265

U.S. Department of Energy, Monthly Energy Review, July 1987.

U.S. Department of Energy, Future Atmospheric Carbon Dioxide Scenarios and Limitation
Strategies, Park Ridge, NJ: Noyes Publications, 1986.
Weinberg, A. M.: 1972 (April), 'Science and Trans-Science', Minerva.
World Bank, World Development Report 1987, New York: Oxford University Press, 1987.
Xu Shoubo: 1988 (April 28), Remarks by Professor Xu Shoubo, Vice Director, Scientific Commit-
tee, Technical Economics Research Institute, Chinese Academy of Social Science, at 'Work-
shop on Developing Country Energy Strategies: Implications for the Greenhouse Problem',
U.S. Environmental Protection Agency, Washington, D.C.
Year 2000 Research Group: 1986 (March 6), China Report, Economic Affairs: China in the Year
2000, Beijing: Science and Technology Documents Publishers, 1984, translated from Chinese
by the Foreign Broadcast Information Service, U.S. National Technical Information Service,
Arlington, VA, JPRS-CEA-86-023.
Zhao, Zong-ci and Kellogg, W.H.: 1988, 'Sensitivity of Soil Moisture to Doubling of Carbon
Dioxide in Climate Model Experiments: Part II: The Asian Monsoon Region', J. Climate 1,
367-378.

(Received 3 August, 1987; in revised form 19 August, 1988).