v>EPA
United States
Environmental Protection
Agency
Office of Air and Radiation
Washington D.C. 20460
EPA 400/1-87/001F
December 1987
Assessing the Risks of
Trace Gases That Can
Modify the Stratosphere
Volume VI:
Technical Support Documentation
Production Projections
<**^fWlll'
-------
Assessing The Risks of Trace Gases
That Can Modify The Stratosphere
Volume VI: Technical Support Documentation
Production Projections
Senior Editor and Author: John S. Hoffman
Office of Air and Radiation
U.S. Environmental Protection Agency
Washington, D.C. 20460
December 1987
U*5. EnTironmental Protection Agency
Region 5, Library (5PL-16)
,;,-.; S. Dearborn Street, Boom 1670
Chicago» ~IL 60604
-------
APPENDIX C
TABLE OF CONTENTS
TAB
Probabilistic Projections of Chlorofluorocarbon Consumption. 1
by William D. Nordhaus (Yale University) and Gary W. Yohe
(Wesleyan University) (1986)
Scenarios of CFG Use: 1985-2075. by Michael J. Gibbs (ICF 2
Incorporated) (1986)
Product Uses and Market Trends for Potential Ozone Depleting 3
Substances 1985-2000. by James K. Hammitt, et al. (RAND
Corporation) (1986)
Joint Emission Scenarios for Potential Ozone Depleting 4
Substances, by Frank Camm, et al. (RAND Corporation) (1986)
-------
PROBABILISTIC PROJECTIONS OF CHLOROFLOUROCARBON
CONSUMPTION
Dr. William D. Nordhaus
Department of Economics
Yale University
New Haven, CT 06520
Dr. Gary W. Yohe
Department of Economics
Wesleyan University
Middletown, CT. 06457
June 1986
-------
Recent scientific evidence indicates that significant ozone depletion
is likely to occur if atmospheric concentrations of chloroflourocarbons
(CFCs) continue to grow at current rates. The precise health and welfare
ramifications of this depletion are not yet completely understood, but the
potential for widespread increases in the incidence of skin cancer, along
with damage to animals and crops, and dramatic climatic change is now
apparent. Increased concentrations are a manmade phenomenon, the result of
increased production and emission of a bevy of fluorocarbons over the past
40 years. Reductions in the emissions of these compounds are possible, but
can be realistically expected only if the cost of the required regulatory
activity can be justified in terms of avoiding expensive environmental
damage.
Assessing the value of a policy response to the potential environmental
consequences of CFC emissions is thus a vital undertaking, but it is
inextricably linked with forecasting CFC emissions into the next century.
It is in the next century that most of the potential problems linked with
CFC emissions are likely to occur, but it is in the next few decades that
policy designed to avoid those problems would be most effective and least
costly. The need for long term forecasting is thus clear, but so too is the
observation that such forecasting is dominated by uncertainty. In the case
of CFC's, rates of emissions beyond the year 2000 will certainly depend upon
circumstances that cannot be accurately predicted.
If the appropriate decisions are to be made in dealing with the
potentially harmful effects of emissions, then, a balance of future risks
and costs must be struck. Given the impossibility of conducting policy
debate with any degree of confidence on the basis of one, two, or ten
-------
alternative scenarios, this balance must be weighed in the context of not
only the estimates of most likely future emissions, but also
probabilistically weighted ranges of other possible trajectories that
deviate from any specified "best guess". Only then can policymakers survey
the full menu of policy options --a menu that may include direct and
immediate intervention, to be sure, but might also include contingency
intervention schemes to be triggered if future scientific findings indicate
that the emissions or health risks are more serious than is now thought.
Despite some significant effort in the past, the literature has not yet
provided policymakers with the evidence required either to evaluate the
relative value of options or to legitimize any but the simplest of options
listed there.
The earliest studies undertaken under the auspices of the National
Academy of Sciences produced estimates of future emissions of CFC's by
simply projecting constant emissions forward from a specified date [see,
e.g., NAS 1976 and 1979]. While these studies helped bring CFC emissions to
the attention of the world scientific community, they were unable to provide
an accurate understanding of the sources of the various forecasts, the
uncertainties surrounding those forecasts, or the relative social values and
costs of the potential policy alternatives. A second generation of
analyses overcame some of the shortcomings of the initial exercises by
constructing precisely articulated scenarios of how the future might unfold
[e.g., Gibbs, ICF, 1986 and Quinn et. al., RAND, 1986]. Each traced future
emissions trajectories from well defined models based upon sets of
explicitly recorded assumptions and statistical analyses. While these
studies have been successful in identifying interesting "what if" scenarios,
-------
they have still been nonprobabilistic in nature. They continue, therefore,
to fail to provide policymakers with any quantified notion of the likelihood
of the specified scenarios.
The present study will try to provide the next step by explicitly
incorporating uncertainty into the modeling exercise so that more definite
estimates of likelihoods can be associated with possible future trajectories
of CFC emissions. The technique to be employed, called probabilistic
scenario analysis, was designed by Nordhaus and Yohe [1983] to investigate
the problem of forecasting carbon dioxide emissions and concentrations from
an economically consistent model of future energy markets. It identifies
the most important sources of uncertainty in our present understanding of
future CFC emissions, examines the current knowledge and disagreement about
these variables and parameters, specifies a range of values that they might
assume with specified relative frequency, and incorporates both the ranges
and their relative frequencies into a simulation or Monte Carlo exercise.
The focus is not, therefore, on resolving uncertainty. It is, instead, on
representing existing uncertainty as accurately as possible and integrating
that uncertainty explicitly and consistently into the analysis. The results
not only provide a "best guess" trajectory for CFC emissions through the
middle of the next century, but also suggest a set of alternative
trajectories and associated probabilities that quantify the range of
outcomes that is consistent with the current state of knowledge and
ignorance.
Section I describes the analytical framework within which the
probabilistic scenario analysis is conducted; coverage of the details of the
modeling has been relegated to supporting appendices. Section II reviews
-------
the judgmental values assigned to the major random variables identified in
Section I. Section III presents the results of the scenario analysis
assuming only the regulatory impetus on reducing aerosol consumption that
produced the ban in the United States nearly a decade ago. Section IV
follows with an extension of the regulatory environment to include the
indefinite continuation of the voluntary restrictions now imposed on aerosol
consumption by various countries across the rest of the world. Comparisons
of the aerosol consumption distributions produced in Sections III and IV can
therefore provide some insight into the possible effect of continuing
voluntary regulatory initiatives outside the United States.
I. The Analytical Framework.
This study analyses the future worldwide emissions of four separate
"CFC commodities": CFC-11 aerosols, CFC-12 aerosols, CFC-11 nonaerosols, and
CFC-12 nonaerosols. Each is derived from the sum of CFC consumption in the
United States and the rest of the world, but the models used to develop the
individual forecasts have different empirical foundations. Because the
models for both components of the four CFC commodities are nonetheless
identical in structure, it is convenient to describe that structure. It is
also convenient to focus initial attention on the modeling that produced the
consumption trajectories, because their derivation is the most elaborate
part of the analysis. CFC emissions flow in part from consumption through a
straightforward linear allocation between emissions from current consumption
and deposits into a CFC bank. The rest of current emissions are the result
-------
of leakage from the existing CFG bank and therefore also depend upon (past)
consumption. To describe the modeling that produced the consumption
scenarios is thus to describe the heart of the analytical framework.
The central uncertainty in projecting future CFC emissions can be
traced to the imprecision with which we understand the demand for CFC
consumption. The general process employed here to model that demand is
illustrated in Figure 1-1. Four separate factors, taken in this study to be
exogenous and uncertain, directly effect the demand for consumption:
technological developments including the development of new products and
processes; the level of population and its distribution across different
regions of the globe; the level of per capita income and its distribution
across different regions of the globe; and the regulatory environment.
Depending upon the demand schedule for CFC consumption, each of these four
plays a role in tracing out the path of future consumption.
The approach taken here to model CFC demand is based upon four major
premises. First of all, most of the observed history of CFC use in the
United States mirrors the experience of a rapidly developing technology in
the early phase of its growth cycle. It should not, however, be expected
that the rapid growth rates of the last two or three decades will continue
into the future as the CFC-using technologies run their courses to maturity
and market saturation in the United States. In our demand modeling,
therefore, we recognize that the growth phase for CFCs has come to an end
(or will do so in the near future), and that the demand patterns for CFCs
are likely to resemble more conventional products in subsequent years.
Secondly, it is assumed that CFC use is likely to be determined by the
presence or absence of major new CFC-using products along with the income
-------
LEVEL AND DISTRIBUTION
OF POPULATION
LEVEL AND DISTRIBUTION
OF INCOME
TECHNOLOGICAL
DEVELOPMENTS
a DEMAND FOR \
="C CONSUMPTION
TON 1
-A
REGULATION
Figure 1-1 Steps in the Generation of CFC Consumption and Emission
-------
elasticity of current CFC-using products. If the inherent possibilities for
major new CFC-using products were to be low because the potential for
regulation inhibited research and development into new CFC uses, then we
should expect that the volume of CFC use per unit of output in the United
States would remain stable or perhaps even decline. If, on the other hand,
major new CFC-using technologies were developed, then, at least for
nonaerosols, the growth in CFC use could continue into the future.
In addition, future trends in CFC use will be significantly affected by
the extent to which current CFC-using products are ones that tend to be more
or less extensively used as peoples' incomes rise (or, technically,
depending upon whether the income elasticities of CFC-using products are
greater or less than unity). If CFC-using products are extensively
contained in CFC-using products that appeal to high-income consumers (like
automobiles or air conditioning), then CFC use will tend to rise more
rapidly than income. Conversely, if CFCs are associated with products like
food that tend to take a lower share of income as income rises, then CFC use
will tend to rise relatively slowly.
In estimating CFC use outside the United States, finally, we view
development as a process through which, abstracting from regulatory
constraints, technologies that have emerged in the most advanced countries
(particularly the United States) are likely to be increasingly adopted by
less advanced countries. Put differently, if and when a particular country
is able to attain the living standards prevailing in the United States, then
we expect that country will exhibit consumption patterns similar to those
supported by today's American consumer --to buy refrigerators and air
conditioners, to purchase deodorants and insecticides, to employ plastic
-------
foams and sprays, and so on. When we project the consumption patterns of
the rest of the world, then, we begin with the assumption that CFC use will
be driven, as have many other commodities, by incomes, prices, and available
technologies.
Figure 1-2 illustrates more specifically the determinants of U.S.
consumption over time. A logistics equation defines a trajectory for
consumption per constant dollar of GNP (locus AB) that asymptotically
converges to an asymptotic CFC frontier (locus CF). The notion behind this
frontier is that the growing technological maturity in CFC markets dictates
gradual movement toward a path of slow growth or decline in the income
intensity of CFC demand (the ratio of CFC use per dollar of real GNP) rather
than a trajectory of immediate convergence. CFCs first came into use in the
1930's, and the widespread introduction of products over the next 30 years
produced an extremely rapid rate of growth. As the potential market became
saturated, however, the growth rate slowed. This abbreviated history is
depicted by the left half of locus AB in Figure 1-2 with the early 1980's
appearing near an inflexion point in the logistics curve. Eventually, if
all markets are saturated and if no new uses are discovered, the actual CFC
frontier will approach some potential CFC frontier like locus CF.
The frontier locus CF depicts the usage pattern when CFC products
ultimately and completely penetrate every available submarket; that is, once
the introduction of the new technologies has been played out. The potential
frontier is thus determined by the level of per capita income in the United
States, the prices of CFC-using products, and climatic, cultural, and other
miscellaneous factors. As drawn in Figure 1-2, the frontier has a positive
slope, but that need not be the case. The slope would be positive if (1)
-------
CFC use/GNP
potential
1950
1980
time
Figure 1-2 Illustration of the Dynamic Logistics Curve for the
U.S. Consumption of CFCs Approaching a Positively
Sloped CFC Frontier
-------
new products and processes developed over the coming decades were CFC-
intensive, (2) CFC-using products displayed income elasticities greater than
unity, or (3) the prices of CFC-using products and services were declining
relative to other commodities. The CFC-intensity of the economy would tend
to decline, giving the frontier a negative slope, if the reverse of these
conditions were true.
It must be emphasized that we have no a priori conviction whether the
CFC frontier has a positive or negative slope. We simply do not know, at
this point, how CFC-intensities, income elasticities, and relative prices
will fit together to determine the trend in the potential for future demand.
It should be clear, nonetheless, that the inclusion of such a frontier in
the modeling focuses attention on a convenient method of capturing all of
that uncertainty with a device that is commonly employed in the economics
literature (e.g., potential GNP). The frontier is, moreover, a convenient
vehicle with which to incorporate even complicated regulatory options in a
straightforward and uncomplicated way. Finally, the dynamic nature of the
frontier, with its potential for a nonzero slope, represents an important
extension of the usual logistics approach that envisions trajectories
climbing asymptotically to a fixed limit.
A system of equations depicting these loci must be estimated on the
basis of U.S. time series data that shows dramatic reductions in the rate of
growth of CFC consumption since the middle of the 1970's --a slowdown that
poses significant difficulty in interpretation. Recall that it was in the
middle of that decade that increased awareness of the potential for CFC's
contribution to ozone depletion produced an aerosol ban in the United
States. General concern that regulatory efforts would extend to other
-------
fluorocarbons was also voiced at that time and probably led to a reduction
in the vitality of CFC market activity at all levels.
The uncertainty about the effect of regulation leads to another
important issue in forecasting trajectories. Figure 1-2 illustrates a
single potential CFC frontier. In fact, we do not know either the level or
the growth rate (positive or negative) of the CFC frontier. This
uncertainty reflects the general problem of "specification uncertainty" in
estimating statistical models.
Figure 1-3 illustrates this difficulty more clearly; there, a second
set of actual and potential trajectories (labeled A'B' and C'D',
respectively) is drawn. Based on the data at hand, it is impossible to sure
whether the frontier is CF or C'F'; i.e., the specification of the
underlying relationship is, itself, uncertain. This type of specification
uncertainty creates fundamental estimation problems in models like the
present one. The procedure for incorporating specification uncertainty into
this study employs historical CFC end-use data for the United States to
estimate the likelihood of different rates of growth supporting different
potential CFC frontiers. Logistics paths converging to these frontiers are
then estimated, again on the basis of the observed U.S. experience. The
procedure is explained in some detail in Appendix A.
One critical source of uncertainty has thus been identified (the
frontier growth parameter), along with a statistical procedure for
estimating its probability distribution. This procedure provides not only a
distribution of growth rates for the frontier, but also an associated
distribution of initial starting points (like C and C' in Figure 1-3).
Locus C'F' begins, in other words, in 1950 (e.g.) at some level of intensity
10
-------
CFC use/GNP
1950
1980
time
Figure 1-3 Illustration of the Possibility of Alternative Paths for
Actual and Potential CFC use
-------
that is consistent with the estimated rate of growth along C'F', and that
initial boundary point need not be the same as the point C that anchors the
first alternative locus CF.
Consumption per constant dollar of GNP for the U.S. can now easily be
translated into total consumption for the U.S. for any given year simply by
multiplying the number read from the AB loci by real per capita GNP and
population for the year in question. Both the rate of growth of per capita
GNP through the middle of the next century, denoted in most economic work as
labor productivity, and the rate of growth of population over the same
period of time are, however, uncertain. Both multiplicative factors are
thus uncertain or random variables. Population and productivity projections
through 2050 are based only partially upon statistical exercises for which
standard errors can be computed. They are, in addition, based upon
judgmental views of the structure of long term economic and demographic
trends that vary from expert to expert. A procedure has been developed to
extract a representation of the underlying (or "true") distribution of that
uncertainty for such a random variable from disagreement across published
projections. The details of the procedure, developed initially for the
Nordhaus and Yohe contribution to the National Academy of Science's Changing
Climate was refined on the basis of some subsequent work by Nordhaus on
macroeconomic forecasts, and its application to productivity and population
growth rates in the United States is outlined in Appendix B.
With consumption for the U.S. sufficiently specified, turn next to the
rest of the world. The United States has, historically, led the rest of the
world to new uses of CFC's. The experience of the rest of the world can
therefore be described reasonably well by a function that reflects the speed
11
-------
with which the rest of the world adopts U.S. patterns of production and
consumption. A time series for the rest of the world can, in fact, be
constructed from an estimated equation that relates differences in the
relative size of CFC consumption per constant dollar of GNP between the rest
of the world and the U.S. to corresponding differences In the relative size
of per capita GNP. Figure 1-4 shows how such a time series might evolve.
Curves AB and CF are defined as in Figure 1-2; that is, they represent
actual and potential CFC usage in the United States. We know that lower-
income countries in Europe, Africa and Asia currently have significantly
lower CFC use per unit of GNP. This is reflected in the rest of the world
curve shown in Figure 1-4 as DE; it lies everywhere below the U.S. curve AB.
The rest of the world curve gradually moves toward the frontier, but it lags
behind the U.S. curve because incomes outside the U.S. are lower and because
there is a diffusion lag of technology developed and introduced in the
United States.
The rate of convergence is determined jointly by an estimated intensity
factor and by projected rates of growth of per capita GNP in the rest of the
world. It is therefore uncertain along two dimensions. Two more sources of
uncertainty must, in other words, now be noted. First, the intensity
parameter must be estimated from cross sectional data, and the confidence of
its estimation captured by a likelihood distribution; this estimation is
also discussed in Appendix A. Secondly, the rate of growth of productivity
in the rest of the world, another exogenous variable whose distribution is
gleaned from disagreement across published projections in Appendix B, enters
at this point. CFC consumption per dollar of GNP has thus been specified,
but a fourth and fifth source of uncertainty have also been revealed.
12
-------
CFC use/GNP
potential, U.S
actual, rest of world
1950
1980
time
Figure 1-4 Illustration of the Rest of the World Given a U.S.
Frontier
-------
A sixth factor appears quickly when CFG consumption per constant dollar
of GNP is translated into total CFC consumption for the rest of the world.
As before, the consumption per dollar series needs to be multiplied by both
per capita GNP and population. Per capita GNP is already available .from
productivity growth projections employed in the transition from U.S.
consumption to the rest of the world, but the level of population has yet to
be specified for the rest of the world. This sixth uncertainty is
quantified by the procedure of comparing forecast disagreement cited above.
Appendix B records the specifics of this process.
A more detailed summary of the entire process can now be presented
schematically; a representation of the computations required each year for
each sampled combination of uncertain parameter and variable values is
depicted in Figure 1-5. The interaction of population, per capita GNP, and
CFC consumption per constant dollar of GNP produces an estimate for total
U.S. consumption of CFC's in period t, as shown at point A in the upper left
hand corner. Two uncertain exogenous variables and one uncertain parameter
imply that three sources of uncertainty need to be quantified just to cover
the U.S. The middle part of the upper half of the figure shows the
interaction of U.S. consumption per constant dollar of U.S. GNP, real per
capita GNP in the U.S., and real per capita GNP in the rest of the world to
produce an estimate of CFC consumption per constant dollar of GNP for the
rest of the world at point B, A third uncertain exogenous variable and a
second uncertain parameter come into play there. The upper right hand part
of the figure reflects at point C the determination of total CFC consumption
for the rest of the world on the basis of CFC consumption per constant
dollar of GNP in the rest of the world, real per capital in the rest of the
13
-------
Frontier
M
H)
CD
S. Population (t)j
__ — -^
U. S. Productivity (t)j
^^. ^ { |{ ^^
U. S. CFC use per $ (t)
S. Consumption (t)
ROW CFC use per $
XTotal World
(Consumption (t)
"
^^ ^^ ^X^
/OFC Emissions (t))
ROW Consumption (t)
Figure 1-5 Detailed Schematic of the Workings of One Iteration of the Analytical Framework
-------
world, and the population level in the rest of the world. The uncertain
exogenous variable required there is the sixth explicit source of
uncertainty that must be quantified to construct a probabilistic scenario
exercise.
The final stage of the model construction, the part that takes CFC
consumption to CFC emissions, is displayed at point E found toward the
bottom of Figure 1-5. Part of the world's total consumption -- the sum of
the U.S. consumption and the rest of the world accomplished at point D --
leaks into the atmosphere during the period of consumption. This fraction
depends upon the mix of products using the CFC's that are produced in any
one year and the difference between the quantity produced and the quantity
consumed. It is claimed that CFC inventories are currently very low
worldwide because they are difficult and expensive to store. It is also
claimed that the decade-long glut of capacity has dampened trade. The
leakage parameter may not, therefore, be significantly affected by the
production/consumption distortion. It may, however, vary significantly with
product mix in a manner that would make the leakage factor from current
consumption the third uncertain parameter of an emissions model and the
seventh source of uncertainty requiring some quantification.
A second source of emission is leakage from the CFC bank that exists
like a vast reservoir of unreleased fluorocarbons. Each year, the
proportion of CFC's consumed which does not leak into the environment is
deposited through point F in Figure 1-5 into a bank from which it will
eventually leak. The schema for period t must, therefore, display some of
current consumption being deposited through point F into the reservoir bank
at the same time it displays some proportion of the bank inherited from
14
-------
period (t-1) leaking through point E into the environment: as emissions. The
first flow proceeds at a rate equal to one minus the current leakage
parameter just defined; its uncertainty has already been noted. The second
flow combines with direct emissions to produce, for any given year, total
emissions and, given well accepted CFC molecular atmospheric lifetimes,
corresponding atmospheric concentrations.
The grand total for identified sources of quantifiable uncertainty is
8. In the Monte Carlo estimates, we discretized the probability
distributions into high, middle, and low cells with 25 percent, 50 percent,
and 25 percent likelihoods to facilitate the sampling process. As a result,
there are 6561 possible combinations of values for the 4 exogenous variables
and 4 estimated parameters or 6561 potential trajectories for CFC emissions
for each of the four types of CFC's to be studied. Even before emissions
are computed, there are 729 possible combinations of 4 random exogenous
variables and 2 estimated parameters upon which total consumption
distributions will depend. Appendix C records the detailed equations of the
complete model just described.
Finally, we need to discuss the regulatory environment presumed
throughout the analysis. Because we begin every run at the actual 1980
consumption levels in the United States and the rest of the world, the
impact of the U.S. aerosol ban is automatically captured in all of the
results. The ban was, of course, applied only to "nonessential" uses, so
1980 consumption levels were not zero, even in the United States. Aerosol
consumption was, instead, some small fraction of the peak levels of the mid
1970's, and the frontier growth methodology is applied to that quantity in
all cases.
15
-------
Section III reports results for the rest of the world based on
scenarios that envision regulation going no further than the U.S. aerosol
ban. These runs assume, in other words, that the voluntary restraint
presently imposed by the EEC and other countries on themselves will lapse
when economic growth makes those restrictions binding. Section IV goes to
the other extreme and reports results based on scenarios that envision
existing worldwide restraint on aerosol consumption and/or capacity
continuing as presently written for the next seventy years. The reader is
thus provided some well defined regulatory benchmarks against which to
measure expectations of the future demand for aerosol regulation. It should
be noted, too, that nonaerosols are never assumed to be subjected to any
direct regulatory restraint. Association with aerosols may have depressed
their consumption levels over the past ten years, but that possibility would
be captured, if it were true, in the lower statistical estimates of the
frontier growth rates.
II. The Subjective Distributions.
Probabilistic scenario analysis applied to the eight sources of
uncertainty just noted requires that we sample randomly over the judgmental
distributions of each source. It is most convenient, for the purposes of
this sampling, to discretize each distribution into high, middle, and low
cells with 25 percent, 50 percent and 25 percent likelihood weights
attached, respectively. Table II-l records the values assigned to these
cells for the four exogenous random variables identified in the previous
section: population growth in the United States and the Rest of the World
and productivity growth in the United States and the Rest of the World. The
technique with which the subjective distributions for these variables were
16
-------
TABLE II-l
Judgmental Cell Values for
Exogenous Random Variables
LIKELIHOOD
VARIABLES WEIGHT 1980-2000
A.
B.
C.
D.
2000-2020
THEREAFTER
U.S. Population Growth (% per year)
High
Middle
Low
Rest of the World
High
Middle
Low
U.S. Productivity
High
Middle
Low
Rest of the World
High
Middle
Low
25
50
25
0.8
0.7
0.6
Population Growth (X
25
50
25
Growth (% per
25
50
25
Productivity
25
50
25
2.0
1.7
1.3
year)
3.2
2.1
1.0
Growth
2.9
2.0
1.1
0.8
0.4
0.3
per year)
1.5
1.1
0.7
3.2
1.6
0.0
(% per year)
2.8
1.8
0.9
0.15
0.1
0.0
0.9
0.5
0.2
1.7
1.1
0.0
1.7
1.1
0.0
-------
produced is described in some detail in Appendix B. The underlying data are
recorded there, as well.
Table II-2 records the cell values assigned to the two uncertain
parameters that specify CFC demand: the rate of growth of the CFG frontiers
for the United States and the relative intensity elasticities for the rest
of the world. They are based upon the estimates that emerge from the
statistical procedures outlined in Appendix A. Table II-3 finally records
the values assumed by the parameters that take consumption of the four CFC
compounds in any one year into current and future emissions: the
distribution of current consumption between release into the atmosphere as
current emission and deposit into the bank for future emission; the annual
rate of release from the bank, and the annual rate of molecular decay in the
atmosphere.
17
-------
TABLE I1-2
Judgmental Cell Values for
Exogenous Random Parameters
LIKELIHOOD
VARIABLES WEIGHT 1980-2000 2000-2020 THEREAFTER
A. Frontier Growth Rate for CFC-11 Nonaerosols
High .25 2.283 1.706 1.749
Middle .50 -0.304 -0.246 -0.250
Low .25 -2.885 -2.214 -2.250
B. Frontier Growth Rate for CFC-12 Nonaerosole
High .25 2.518 2.0 2.0
Middle .50 -0.729 0.0 0.0
Low .25 -2.647 -2.0 -2.0
C. Frontier Growth rate for CFC-11 Aerosols
High .25 2.451 2.0 2.0
Middle .50 -0.043 0.0 0.0
Low .25 -2.590 -2.0 -2.0
D. Frontier Growth Rate for CFC-12 Aerosols
High .25 2.451 2.0 2.0
Middle .50 -0.095 0.0 0.0
Low .25 -2.596 -2.0 -2.0
E. Relative Intensity Elasticity: Rest of the World for CFC-11 Nonaerosols
High .25 1.62 1.62 1.62
Middle .50 1.27 1.27 1.27
Low .25 0.92 0.92 0.92
F. Relative Intensity Elasticity: Rest of the World for CFC-12 Nonaerosols
High .25 1.43 1.43 1.43
Middle .50 1.15 1.15 1.15
Low .25 0.87 0.87 0.87
G. Relative Intensity Elasticity: Rest of the World for CFC-11 Aerosols
High .25 1.82 1.82 1.82
Middle .50 1.30 1.30 1.30
Low .25 0.78 0.78 0.78
H. Relative Intensity Elasticity: Rest of the World for CFC-12 Aerosols
High .25 1.31 1.31 1.31
Middle .50 0.93 0.93 0.93
Low .25 0.55 0.55 0.55
-------
TABLE II-3
Judgmental Cell Values
for Emissions Parameters
LIKELIHOOD
VARIABLE WEIGHT VALUE
A. Fraction of Consumption emitted Directly:
CFC-11 Nonaerosols 1.0 0.1
CFC-12 Nonaerosols 1.0 0.1
CFC-11 Aerosols 1.0 1.0
CFC-12 Aerosols 1.0 1.0
G. Fraction of Bank Emitted:
CFG High .25 .85
CFC Middle .50 .80
CFC Low .25 .75
C. Annual Rate of Atmospheric Decay
CFC-11 Nonaerosols 1.0 0.01
CFC-11 Nonaerosols 1.0 0.0143
CFC-12 Aerosols 1.0 0.01
CFC-12 Aerosols 1.0 0.043
-------
III. Results
In reporting the results of this analysis, we show not only the global
consumption (production), emissions, mass in the banks, and concentrations
for the identified CFC compound, but also a number of important ingredients:
the division of total world production of CFCs between the U. S. and the
rest of the world, GNP levels for the U. S. and the Rest of the World,
population for the U.S. and the rest of the world, and other key driving
variables. Recall, particularly in reading the aerosol results recorded in
this section, that they include only the effects of the U.S. aerosol ban.
Voluntary restrictions presently in place in various countries around the
world are reflected explicitly in the results reported in Section IV.
Nonaerosols are assumed to face no formal restraint. Finally, the
concentration and bank volume trajectories are nominally based on estimates
of the 1980 levels, but the long term results are almost completely
insensitive to their initial specification; virtually all of the CFC's in
the banks and in the atmosphere in the year 2020 and beyond will have been
produced since 1980.
Table III-l shows the base or "most likely" cases obtained by setting
each of the uncertain random variables equal to its middle or most likely
value. The remaining tables present the probabilistic results based on 200
randomly selected runs. The runs were chosen by drawing one of the three
values for each of the uncertain variables and parameters from a random
number generator; the probability of selecting each was proportional to the
corresponding judgmental probability under the assumption that all of the
variables were independently distributed. For example, because there were 7
uncertain variables (the distribution of CFC consumption between immediate
-------
Table II1-1
Base Case Seceoarioa
Major Driving Variables:
U.S. POPULATION U0*«9)
ROW POl>ULATION (10"9)
US LABOR PRODUCTIVITY (1985$ 10««3/CAP«YR)
RW LABOR PRODUCTIVITY (1985$ 10««3/CAP«YR)
U.S CROSS OUTPUT (1985* 10"12/YR)
ROW CROSS OUTPUT (1984$ 10"12/YR)
WORLD CROSS OUTPUT (1985$ 10"12/YR)
CFC-11 Konaerosols:
RATIO OF CFC/CKP (US) (10"3 MT/YR)
US CFC CONSUMPTION (10"3 MT/YR)
ROW CFC CONSUMPTION (10**3 MT/YR.)
WORLD CFC CONSUMPTION (10*»3 MT/YR)
WORLD CFC BANK (10««3 MT)
WORLD CFC BUSSIONS (10"3 MT/TR)
WORLD CFC CONCENTRATIONS U0"3 MT)
1980
1990
2000
2010
2020
2030
2040
193.316 264.650
357.122 464.330 589.657
2050
0.225
4.226
15.150
2.520
3.409
10.550
14.058
19.836
67.616
125.700
0.241
5.009
18.690
3.078
4.510
15.418
19.928
19.504
87.969
176.680
0.258
5.884
22.885
3.748
5.896
22.054
27.951
19.032
112.222
244.901
0.269
6.608
26.991
4.496
7.260
29.711
36.970
18.593
134.983
329.347
0.279
7.310
31.437
5.327
8.761
38.939
47.701
18.153
159.041
430.616
0.282
7.731
35.268
5.988
9.958
46.294
55.251
17.710
175.346
500.719
0.285
8.127
39.369
6.684
11.227
54.326
65.553
17.274
193.937
573.142
0.288
8.544
43.947
7.461
12.658
63.752
76.410
16.848
213.270
656.005
677.065 767.078 869.275
689.193 1.051.439 1.438.844 1,893.235 2,435.381 2.888.110 3,287.048 3,726.254
148.134 229.779 315.072 415.269 535.140
637.397 725.880
822.907
1,935.351 3,618.248 5911.371 8.882.086 12,631.320 17,115.297 22,042.629 27,384.535
CFC-12 Nooaerosols:
RATIO OF CFC/GKP (US) (10««3 MT/YR)
US CFC CONSUMPTION (10««3 MT/YR)
ROW CFC CONSUMPTION (10**3 MT/YR)
WORLD CFC CONSUMPTION (10**3 MT/YR)
WORLD CFC BANK U0**3 MT)
WORLD CFC EMISSION (10"3 MI/YR)
WORLD CFC CONCENTRAIONS (10««3 MT)
37.647
128.324
120.800
249.124
883.011
189.612
35.427
159.781
162.697
322.478
1,303.290
285.514
33.448
197.224
218.350
415.574
1.697.653
372.468
33.261
241.461
298.276
539.738
2,201.117
482.814
33.281
291.582
398.888
690.470
2,843.410
624.546
33.288
331.471
475.424
806.894
3,419.687
754.063
33.291
373.471
557.960
931.722
3,964.472
874.646
33.292
421.425
654.794
1.076.220
4.580.465
1,010.593
2.749.781 4,699.805 7,195.145 10,281.656 14,160.613 18,827.820 24,003.590 29,690.176
CFC-11 Aerosols:
RATIO OF CFC/GNP (US) (10*«3 MT/YR)
ui CfC CONSUMPTION (10**3 MT/YR)
ROW CFC CONSUMPTION (10««3 MT/YR)
WORLD CFC CONSUMPTION (10**3 MI/YR)
WORLD CFC BANK U0**3 MT)
WORLD CFC EMISSIONS (10*«3 MT/YR)
WORLD CFC CONCENTRATIONS (10«*3 MT)
CFC-12 Aerosols:
RATIO OF CFC/GNP (US) U0«*3 MT/YK)
US CFC CONSUMPTION (10*«3 MT/YR)
ROW CFC CONSUMPTION (10**3 MT/YR)
WORLD CFC CONSUMPTION (10*«3 MI/YR)
WORLD CFC BANK U0**3 MT)
WORLD CFC EMISSIONS <10*«3 MT/YR)
WORLD CFC CONCENTRATIONS (10*«3 MT)
24.062
4.101
111.900
116.001
0.000
116.001
2,506.029
31.695
5.402
178.800
184.202
0.000
184.202
24.282
5.476
161.371
166.847
0.000
166.847
3,633.926
31.797
7.171
257.287
264.457
0.000
264.457
24.314
7.168
229.493
236.661
0.000
236.661
5,239.586
31.690
9.343
364.936
374.278
0.000
374.278
24.348
8.838
316.515
J25.353
0.000
325.353
7,456.504
31.716
11.512
499.856
511.368
0.000
511.368
24.365
10.673
424.388
435.061
0.000
435.061
10,442.164
31.736
13.902
665.987
679.889
0.000
679.889
24.371
12.134
505.979
581.113
0.000
518.113
14,059.691
31.743
15.804
793.429
809.234
0.000
809.234
24.373
13.682
593.827
607.409
0.000
607.509
18,135.649
31.746
17.821
931.178
948.999
0.000
9A8.999
24.374
15.427
696.887
712.313
0.000
712.313
22,756.617
31.747
20.093
1,092.786
1,112.879
0.000
1,112.879
3,300.393 5,057.617 7,416.750 10,628.937 14,891.117 19,970.742 25,602.395 31.913.781
-------
release into the atmosphere and deposit into the bank was not included as a
random variable in the probabilistic runs), the most likely run had .0078 -
.5 chance of being chosen.
Table III-2 presents the mean values of the different variables for the
period 1980 through 2050. As can be seen, these differ from the most likely
values shown in Table III-l. More precisely, the mean Levels of CFC
consumption, emission, bank volume, and atmospheric concentration are all
higher when uncertainty is allowed than when uncertainty is ignored. The
reason for this difference can be traced to the nonlinear nature of the
demand model. Location of the mean above the most likely scenario can occur
only if the extreme trajectories on the high side deviate further from the
best guess than do the extreme runs on the low side. Further evidence of
this skewness will be drawn from tabulations of 25th and 75th percentile
boundaries provided below.
Table III-3 records the rates of growth of the different variables over
the next seven decades. These show that CFC production grows faster than
GNP for both aerosols and non aerosols over the near term. By the middle of
the next century, though, consumption increases slow, but to rates that
still exceed the rate of growth of GNP. Remember that these growth rates
emerge from simulations that ignore the possibility of increased regulatory
activity.
Table III-4 shows the quartile projections of the estimates of CFC
production along with the other critical variables. The 25th and 75th
percentile paths show, respectively, the trajectory below which the outcome
is thought to lie with probability one-quarter -nd above which the outcome
is thought to lie with probability one-quarter. Put differently, the
-------
Table III-2
Mean Eatlaaces Based OD Probabilistic Scenario
1980
1990
2000
2010
2U20
2030
2040
2050
Major Driving Variables:
US POPUI-ATION U0««9)
ROW POPULATION (10«*9)
US LABOR PRODUCTIVITY (198Si 10*«3/CAP«YK)
RW LABOR PRODUCTIVITY (1985* 10«*3/CAP«YR)
US CROSS OUTPUT (1985* 10*«12/Wl)
ROW CROSS OUTPUT U984J 10*«12/YR)
WORLD GROSS OUTPUT (1985i 10««12/YR)
CFC-11 Nouaerosola:
RATIO OF CFC/CNP (US) (10"3 KT/YR)
US CFC CONSUMPTION (10«*3 KT/YR)
ROW CFC CONSUMPTION (10**3 KT/YR)
WORLD CFC CONSUMPTION (10"3 KI/YR)
WORLD CFC BANK (10"3 KT)
WORLD CFC EMISSIONS U0««3 KT/YR)
WORLD CFC CONCENTRATIONS (10"3 KT)
CFC-12 Nooaero»ol«:
RATIO OF CFC/CNP (US) U0**3 XT/ML)
US CFC CONSUMPTION (10"3 KT/YR)
ROW CFC CONSUMPTION (10*«3 KT/YR)
WORLD CFC CONSUMPTION (10"3 KT/YR)
WORLD CFC BANK (10"3 KI)
WORLD CFC EMISSION (10««3 KT/YR)
WORLD CFC CONCENTRAIONS U0"*3 HT)
0.225
4.226
15.150
2.520
3.409
10.649
14.058
19.836
67.614
125.699
193.313
687.732
149.581
1,936.767
37.645
128.321
120.797
249.121
881.141
191.463
0.241
5.015
18.915
3.087
4.568
15.481
20.049
20.090
91.881
182.803
274.686
1,084.663
233.000
3,624.866
37.772
172.787
174.039
346.826
1,377.701
296.206
0.
5.
23.
3.
6.
22.
28.
20.
127.
277.
405.
1,603.
343.
6,059.
39.
239.
263.
503.
2.000.
428.
25«
904
596
786
092
358
449
793
288
910
195
191
462
812
127
732
400
124
347
816
0.270
6.665
28.779
4.603
7.783
30.694
38.477
21.748
171.174
431.751
602.921
2,378.147
509.032
9.594.180
42.056
331.542
414.791
746.321
2,947.394
630.988
0.
7.
34.
5.
9.
41.
50.
23.
270
421
879
539
807
137
944
069
230.565
682.
912.
3,601.
771.
14,855
46.
461.
664.
1,125.
4,445.
951.
259
824
833
112
.895
033
098
279
379
590
686
0
7
39
5
11
49
61
24
290
962
1,252
5,090
1,093
22,507
51
598
952
1,551
6,295
1,351
.281
.913
.874
.267
.360
.648
.007
.872
.039
.655
.689
.520
.563
.977
.176
.228
.858
.099
.930
.506
1,
1,
7.
1,
32
1,
2,
8,
1,
0.287
8.395
45.298
7.050
13.031
59.264
72.295
27.201
366.555
369.580
736.134
023.227
508.082
,905.691
57.700
779.570
377.317
156.903
722.164
870.871
0.290
8.901
51.570
7.954
14.981
70.986
85.968
30.123
470.212
2,000.110
2,470.333
9,902.477
2,125.121
47,272.129
65.869
1,030.872
2,040.773
3,071.668
12,323.117
2,640.922
2,751.606 4,734.773 7,535.559 11,544.707 17,502.113 26,154.359 37,915.098 54,195.031
CFC-11 Aeroioli:
RATIO OF CFC/CNP (US) (10"3 KT/YR)
US CFC CONSUMPTION (10*«3 KI/YR)
ROW CFC CONSUMPTION (10"3 MT/YX)
WORLD CFC CONSUMPTION (10"3 MT/TR)
WORLD CFC BANK (10«*3 KT)
WORLD CFC EMISSIONS (10**3 KT/YR)
WORLD CFC CONCENTRATIONS (10*«3 KT)
CFC-12 Aerocola:
RATIO OF CFC/GNP (US) (10*«3 KT/YR)
US CFC CONSUMPTION (10««3 KT/YR)
ROW CFC CONSUMPTION (10"3 KT/YR)
WORLD CFC CONSUMPTION (10««3 KT/YR)
WORLD CFC BANK (10««3 KT)
WORLD CFC EMISSIONS U0*«3 KT/YR)
WORLD CFC CONCENTRATIONS (10««3 HT)
24.061
4.101
111.899
116.000
0.000
116.000
2,506.003
31.693
5.402
178.797
184.200
0.000
184.200
24.982
5.713
166.782
172.495
0.000
172.495
3,654.350
32.632
7.290
270.379
277.669
O.OOO
277.669
26.489
8.108
260.069
268.177
0.000
268.177
5,426.937
34.352
10.019
426.177
436.193
O.OOO
436.193
28
11
416
427
0
427
8,254
36
13
682
696
0
696
.426
.186
.439
.624
.000
.624
.383
.724
.368
.794
.163
.000
.163
30.
15.
680.
695.
0.
695.
12.912
39.
17.
1 , 104 .
1,122.
0.
1,122.
957
470
136
605
000
605
.727
871
982
890
887
OOO
870
34
19
989
1.009
0
1,009
19,949
44
22
1,595
1,618
o
1.618
.230
.958
.654
.613
.000
.613
.750
.002
.829
.984
.817
.OOO
.817
38.
25.
1,452.
1,478.
0.
1,478.
30,030
49.
29.
2,319.
2,348.
0.
2.348.
385
864
441
304
000
304
.828
268
224
579
805
000
805
43.589
34.021
2,189.739
2.223.763
0.000
2,223.763
44,998.930
55.878
38.076
3,452.257
3,490.336
0.000
3,490.336
-------
TABLE III-3
Kates of Growth-Bate Case and Mean Estimates
1980-
1990
1990-
2000
2000-
2010
2010-
2020
2020-
2030
2030-
2040
2040-
2050
Base Case
US POPULATION
ROW POPULATION
US LABOR PRODUCTIVITY
RW LABOR PRODUCTIVITY
US GROSS OUTPUT
ROW GROSS OUTPUT
WORLD GROSS OUTPUT
CFC-11 Nooaerosola:
RATIO OF CFC/CNP (US)
US CFC CONSUMPTION
ROV CFC CONSUMPTION
WORLD CFC CONSUMPTION
WORLD CFC BANK
WORLD CFC EMISSIONS
WORLD CFC CONCENTRATIONS
CFC-12 Nooaerosols:
RATIO OF CFC/CNP (US)
US CFC CONSUMPTION
ROW CFC CONSUMPTION
WORLD CFC CONSUMPTION
WORLD CFC BANK
WORLD CFC EMISSIONS
WORLD CFC CONCENTRATIONS
CFC-11 Aerosols:
RATIO Cf CFC/CNP (US)
US CFC CONSUMPTION
ROW CFC CONSUMPTION
WORLD CFC CONSUMPTION
WORLD CFC BANK
WORLD CFC EMISSIONS
WORLD CFC CONCENTRATIONS
CFC-12 Aerosols:
RATIO OF CFC/CNP (US)
US CFC CONSUMPTION
ROW CFC CONSUMPTION
WORLD CFC CONSUMPTION
WORLD CFC BANK
WORLD CFC EMISSIONS
WORLD CFC CONCENTRATIONS
0.700
1.700
2.100
2.000
2.800
3.700
3.469
0.655
1.610
2.025
1.970
2.680
3.560
3.363
0.430
1.160
1.650
1.820
2.080
2.960
2.797
0.355
1.010
1.525
1.695
1.880
2.705
2.548
0.130
0.560
1.150
1.170
1.280
1.730
1.649
0.100
0.500
1.100
1.100
1.200
1.600
1.530
0.100
0.500
1.100
1.100
1.200
1.600
1.533
-0.169
2.631
3.404
3.141
4.224
4.390
6.257
-0.245
2.435
3.265
2.997
3.137
3.137
4.909
-0.233
1.647
2.963
2.625
2.744
2.761
4.072
-0.240
1.640
2.681
2.389
2.518
2.536
3.521
-0.247
1.033
1.508
1.382
1.705
1.749
3.038
-O.249
0.951
1.351
1.248
1.294
1.300
2.530
-0.250
0.950
1.350
1.251
1.254
1.255
2.170
-0.608
2.192
2.978
2.581
3.893
4.093
5.360
-0.575
2.105
2.942
2.536
2.644
2.659
4.259
-0.056
2.024
3.199
2.614
2.597
2.595
3.570
0.006
1.886
2.907
2.463
2.560
2.574
3.201
0.002
1.282
1.755
1.558
1.845
1.885
2.849
0.001
1.201
1.601
1.438
1.478
1.484
2.429
0.091
2.891
3.661
3.635
0.000
3.635
3.716
0.013
2.693
3.522
3.496
0.000
3.496
3.659
0.014
2.094
3.215
3.183
0.000
3.163
3.528
0.007
1.887
2.933
2.906
0.000
2.906
3.368
0.002
1.282
1.758
1.747
0.000
1.747
2.975
0.001
1.201
1.601
1.592
0.000
1.592
2.546
0.000
1.200
1.600
1.442
1.444
1.445
2.126
0.000
1.200
1.600
1.592
0.000
1.592
2.270
0.032
2.832
3.639
3.616
0.000
3.616
4.029
-0.034
2.646
3.495
3.473
0.000
3.473
3.828
0.008
2.088
3.147
3.121
0.000
3.121
3.598
0.006
1.886
2.869
2.848
0.000
2.848
3.372
0.002
1.282
1.751
1.742
0.000
1.742
2.935
0.001
1.201
1.601
1.593
0.000
1.593
2.484
0.000
1.200
1.600
1.593
0.000
1.593
2.204
-------
TABLE III-3
(Continued)
Rites of Growth-Base Case and Mean EatlMtea
1980-
1990
1990-
2000
20OO-
2010
2010-
2020
2020-
2030
2030-
2040
2040-
2050
Means
US POPULATION
ROW POPULATION
US LABOR PRODUCTIVITY
RU LABOR PRODUCTIVITY
US CROSS OUTPUT
ROW GROSS OUTPUT
WORLD CROSS OUTPUT
CFC-11 Nonaeroaola:
RATIO OF CFC/CNP (US)
US CFC CONSUMPTION
ROW CFC CONSUMPTION
WORLD CFC CONSUMPTION
WORLD CFC BANK
WORLD CFC EMISSIONS
WORLD CFC CONCENTRATIONS
CFC-17 Nonaerosols:
RATIO OF CFC/GNP (US)
US CFC CONSUMPTION
ROW CFC CONSUMPTION
WORLD CFC CONSUMPTION
WORLD CFC BANK
WORLD CFC EMISSIONS
WORLD CFC CONCENTRATIONS
CFC-11 Aerosols:
RATIO OF CFC/GNP (US)
US CFC CONSUMPTION
ROW CFC CONSUMPTION
WORLD CFC CONSUMPTION
WORLD CFC BANK
WORLD CFC EMISSIONS
WORLD CFC CONCENTRATIONS
CFC-12 Aerosols:
RATIO OF CFC/GNP (US)
US CFC CONSUMPTION
ROW CFC CONSUMPTION
WORLD CFC CONSUMPTION
WORLD CFC BANK
WORLD CFC EMISSIONS
WORLD OT
0.707
1.712
2.220
2.029
2.927
3.7*1
3.550
0.667
1.632
2.211
2.041
2.879
3.676
3.499
0.463
1.211
1.986
1.954
2.451
3.169
3.019
0.386
1.074
1.922
1.850
2.311
2.929
2.807
0.130
0.643
1.339
1.235
1.470
1.880
1.803
0.096
0.590
1.275
1.178
1.373
1.770
1.698
0.096
0.596
1.297
1.206
1.395
1.805
1.732
0.127
3.067
3.745
3.513
4.556
4.432
6.268
0.344
3.260
4.189
3.887
3.907
3.880
5.139
0.449
2.962
4.405
3.974
3.943
3.934
4.595
0.590
2.979
4.576
4.148
4.151
4.153
4.372
0.753
2.295
3.443
3.165
3.459
3.494
4.155
0.895
2.341
3.526
3.264
3.218
3.214
3.798
0.034
2.975
3.652
3.309
4.470
4.364
5.427
0.352
3.275
4.144
3.720
3.7M
3.700
4.647
0.722
3.242
4.541
3.943
3.876
3.863
4.266
0.904
3.299
4.709
4.107
4.110
4.109
4.161
1.059
2.604
3.608
3.208
3.480
3.507
4.017
0.376
3.315
3.991
3.968
0.000
3.968
3.772
0.586
3.501
4.443
4.413
0.000
4.413
3.955
0.706
3.219
4.708
4.666
0.000
4.666
4.194
0.853
3.242
4.906
4.865
0.000
4.865
4.475
1.005
2.547
3.751
3.725
0.000
3.725
4.350
1.200
2.648
3.684
3.297
3.260
3.252
3.713
1.146
2.592
3.836
3.813
0.000
3.813
4.090
0.292
2.997
4.136
4.104
0.000
4.104
4.132
0.514
3.179
4.550
4.517
0.000
4.517
4.233
0.668
2.884
4.713
4.675
0.000
4.675
4.397
0.822
2.965
4.813
4.781
0.000
4.781
4.586
0.9086
2.387
3.677
3.658
0.000
3.658
4.367
1.130
2.470
3.739
3.722
0.000
3.722
4.051
1.021
2.490
3.787
3.527
3.436
3.430
3.623
1.324
2.794
3.932
3.535
3.456
3.447
3.572
1.271
2.741
4.105
4.083
0.000
4.083
4.044
1.259
2.646
3.976
3.961
0.000
3.961
3.973
-------
TABLE III-*
25th .nd 75th per«"tll« Scea.rlo.
Based on Prob.bllltl.tlc Seeiurlos
1990
2000
2010
2020
2030
2040
2050
Major Driving Variable.:
251 Point. fro« Hlghe.t
U.S. POPULATION (10*«9)
ROW POPULATION (10**9)
US LABOR PRODUCTIVITY (1985$ 10"3/CAP*YR)
RW LABOR PRODUCTIVITY (1985$ io«*3/CAP*tt)
U.S CROSS OUTPUT (198Si 10««12/YR)
ROW CROSS OUTPUT (1984$ 10**12/YR)
WORLD CROSS OUTPUT (1985$ 10**12/tR)
25Z Point, fro* Love.t
US POPULATION (10««9)
ROW POPULATION (10**9)
US LABOR PRODUCTIVITY (1985$ 10*«3/CAP«YR.)
RW LABOR PRODUCTIVITY (1985$ 10*«3/CAP«Y8.)
US GROSS OUPUT (1985$ 10««12/YR)
ROW CROSS OUPUT (1985$ 10*«12/YR).650
WORLD GROSS OUTPUT (1985$ 10"12/YR)
0.263
5.162
20.863
3.078
4.985
15.887
20.922
0.241
5.009
18.690
3.078
4.465
14.BL3
19.323
0.280
6.257
28.731
3.748
7.256
23.454
30.856
0.258
5.884
22.885
3.748
5.780
20.359
26.255
0.295
7.306
39.567
4.496
10.328
32.852
43.494
0.269
6.608
26.991
4.496
7.045
26.351
33.611
0.301
8.413
53.276
5.327
14.265
44.814
59.661
0.279
7.310
31.437
5.327
8.418
33.232
41.993
0.306
9.260
64.103
5.988
17.216
55.452
73.550
0.282
7.731
35.268
5.988
9.472
38.302
48.259
0.310
10.133
75.981
6.684
20.406
67.729
89.397
0.265
8.127
39.369
6.684
10.573
43.619
54.846
11.067
90.060
7.461
24.187
82.724
108.665
0.288
8.544
43.947
7.461
11.802
49.675
62.333
CFC-11 Noa.ero.ol.:
25! Point, fro* Hlghe.t
RATIO OF C/C/GW (US) (10"3 KI/YR)
US CFC CONSUMPTION (10»*3 MT/YR)
RC*. CFC CONSUMPTION (10««3 MT/YR)
WORLD CFC CONSUMPTION (10"3 MT/YR)
VOiC-L CFC BANK ac««i HI)
WORLD CFC EMISSION (10«*3 MT/YR)
WORLD CFC CONCENTRAIONS (10*«3 MT)
25.263
103.098
216.726
311.733
1,236.624
254.521
3,768.578
31.627
151.766
369.757
506.536
1,915.576
413.702
6,581.363
37.797
206.053
612.346
784.184
3,081.874
650.924
44.887
285.687
1,007.599
1,176.177
4,694.414
994.623
11,180.367 18,018.219
53.457
341.956
1,354.829
1,617.706
6,737.695
1,395.858
27,858.199
63.680
401.323
1,777.063
2,215.491
9,239.391
1,902.741
41,106.008
75.854
470.970
2,330.753
3,045.602
12,442.797
2.598.740
58,880.111
251 Point, from Lowe.t
RATIO OF CFC/GNP (US) U0**3 MT/YR)
US CFC CONSUMPTION (10'*3 MT/YR)
ROW CFC CONSUMPTION (10««3 MT/YR)
WORLD CFC CONSUMPTION (10*«3 MT/YR)
WORLD CFC BANK (10««3 MT)
WORLD CFC EMISSIONS (10*«3 MT/YR)
WOKLD CFC CONCENTRATIONS (10**3 MT)
17.274 16.848
95.406 90.303
193.402 188.634
427.685 446.512
1,941.073 2,082.438
412.398 441.535
3,473.714 5,450.258 7,568.594 10,062.211 12,712.559 15,189.402 17,991.176
19.504
78.806
1*7.620
230.760
903.568
208.710
19.032
89.387
169.343
274.975
1,140.644
251.501
18.593
94.415
179.709
330.995
1,372.464
303.486
18.153
98.731
198.749
390.663
1,666.671
361.648
17.710
98.218
196.861
406.863
1,849.729
394.130
-------
TABLE III-*
(Continued)
25th tad 75th Percentlle Scenario*
Baaed on Probabllltlstlc Scenario*
1990
2000
2010
2020
2030
2040
2050
CFC-12 Nonaerosols:
231 Points from Highest
RATIO OF CFC/CNP (US) (10**3 MT/YR)
US CFC CONSUMPTION (10**3 MI/YR)
ROW CFC CONSUMPTION (10**3 KT/YR)
WORLD CFC CONSUMPTION (10««3 MT/YR)
WORLD CFC BANX (10««3 MT)
WORLD CFC EMISSIONS (10"3 KT/YR)
WORLD CFC CONCENTRATIONS (10**3 MT)
251 Points fro* Lowest
RATIO OF CFC/CNP (US) (10**3 KT/YB.)
US CFC CONSUMPTION (10**3 MT/YR)
ROW CFC CONSUMPTION (10"3 KT/YR)
WORLD CFC CONSUMPTION (10"3 MT/YR.)
WORLD CFC BANK (10*«3 KT)
WORLD CFC EMISSIONS (10"3 MT/YR)
WORLD CFC CONCENTRATIONS (10**3 HI)
CFC-11 Aerosols:
25Z Points fro« Highest
RATIO OF CFC/CNP (US) (10«*3 MT/YR)
US CFC CONSUMPTION (10*«3 MT/YR)
ROW CFC CONSUMPTION (10«*3 KT/YR)
WORLD CFC CONSUMPTION (10**3 MT/YR)
WORLD CFC BANK (10««3 KT)
WORLD CFC EMISSIONS U0*«3 KT/tt)
WORLD CFC CONCENTRATIONS (10"3 Ml)
25Z Points froa Lowest
RATIO OF CFC/CNP (US) (10««3 MT/YR)
US CFC CONSUMPTION (10**3 KT/YR)
ROW CFC CONSUMPTION (10«*3 MT/YR)
WORLD CFC CONSUMPTION (10««3 MT/YR)
WORLD CFC BANK (10«*3 KT)
WORLD CFC EMISSIONS (10**3 KT/YR)
WORLD CFC CONCENTRATIONS (10**3 MT)
49.017
200.035
197. 649
406. 578
1,608.779
321.359
4.903.859
35.427
144.575
138.628
291.868
1.153.086
265.582
4.566.410
31.297
6.386
198.482
203.958
0.000
203.958
3.807.616
24.282
4.905
134.384
140.497
0.000
140.497
3,503.788
62
301
322
651
2.448
538
8,251
.846
.57
.568
.229
.530
.702
.828
33.448
160
158
345
1.404
318
6.784
.505
.029
.501
.957
.594
.160
40.068
9
346
355
0
355
6.139
24
5
157
166
0
166
4,655
.614
.312
.553
.000
.553
.672
.314
.710
.610
.609
.000
.609
.246
77
398
545
1.000
3.859
841
13.770
33
171
183
409
1.738
376
9.187
49
13
603
612
0
612
.288
.559
.165
.736
.858
.222
.059
.261
.519
.845
.812
.043
.784
.453
.272
.491
.360
.198
.000
.198
10,213.488
24
6
172
183
0
183
5,898
.348
.202
.506
.693
.000
.693
.242
94.457
523.771
892.341
1,629.301
6.011.305
1.321.166
21.886.211
33.281
181.010
208.512
490.486
2.103.559
449.639
11.772.969
60.222
19.172
1.004.243
1,014.916
0.000
1.014.916
17.174.477
24.365
6.626
195.365
201.608
0.000
201.608
7.181.227
115
.397
642.762
1.220
2.387
8.782
1.971
34.148
33
184
211
522
2,325
501
14,651
73
23
1.387
1.400
0
1.400
27,287
24
6
198
214
0
214
8,488
.446
.655
.383
.746
.199
.288
.616
.656
.164
.642
.927
.266
.574
.528
.192
.137
.000
.137
.613
.371
.758
.280
.583
.000
.583
.150
140
793
1,663
3,372
11.837
2,905
52,692
33
187
211
559
2,656
537
17,508
89
28
1.868
1,883
0
1,883
40,535
24
6
198
207
0
207
9,805
.958
.560
.748
.282
.477
.516
.902
.291
.421
.673
.566
.364
.825
.289
.872
.312
.938
.608
.000
.608
.016
.373
.732
.298
.788
.000
.788
.414
2
4
15
4
78
2
20
2
2
2
58
10
172.172
983.937
,285.084
,602.164
,966.754
,033.716
.302.437
33.292
190.260
223.563
600.251
,813.909
576.768
,500.742
109.774
34.067
,517.841
.541.287
0.000
,541.287
,003.902
24.374
6.533
198.305
220.319
0.000
220.319
,991.805
-------
TABLE III-*
(Continued)
23th mad 75th Ferceatlle Scenarios
Baaed on Probabllltlatlc Scenario*
1990
2000
2010
2020
2030
CPC-12 Aeroaola:
25X Polnta from Hlgheat
RATIO OF CTC/Off (US) (10"3 MT/TR)
OS CFC CONSUMPTION (10**3 MT/TR.)
ROW CTC CONSUMPTION (10"3 MT/TR)
WORLD ere CONSUMPTION (10"3 MT/TR)
WORLD CTC BANK (10*«3 MI)
WORLD Crc EMISSIONS (10**3 MT/TR)
WORLD Cre CONCENTRATIONS (10**3 HT)
25Z Polnta from Loweat
RATIO OP CTC/ONP (OS) (10**3 MT/1H.)
OS Cre CONSUMPTION U0*«3 MT/TR)
ROW Crc CONSUMPTION (10**3 MT/TR)
WORLD Cre CONSUMPTION (10««3 MT/TR)
WORLD Crc BANK (10**3 MI)
WORLD Crc EMISSIONS (10**3 MT/TR)
WORLD CTC CONCENTRATIONS (10«*3 MT)
2040
2050
41.016
8.286
315.540
322.711
0.000
322.711
5.327.379
31.797
6.360
222.848
230.090
0.000
230.090
4,890.773
52.284
12.278
551.745
561.088
0.000
561.088
8.798.227
31.690
7.294
274.189
283.735
0.000
283.735
6,671.895
64.133
16.536
896.992
920.720
0.000
920.720
14,816.363
31.716
7.853
328.231
339.614
0.000
339.614
8,723.535
78.236
22.636
1,505.762
1.509.080
0.000
1.509.080
24.085.199
31.736
8.142
380.385
395.121
0.000
395.121
11.047.801
95.559
27.324
2.061.926
2,064.678
0.000
2.064.678
37,460.031
31.743
8.251
382.926
403.926
0.000
403.926
13,335.801
116.727
32.857
2,770.571
2,836.096
0.000
2,836.096
55.562.504
31.746
8.334
405.643
438.033
0.000
438.033
15,284.191
142.576
40.740
3.853.907
3.944.146
0.000
3,944.146
80,272.687
31.747
8.418
436.901
475.294
0.000
475.294
17.503.582
-------
results suggest that it is equally likely that the value of future CFC
production will lie between the two paths as it is that it will lie outside
their boundary.
A few words may be necessary to explain exactly what this and the
other dispersion measures are intended to convey. At the present time, we
have some information, but only very imperfect information, about the
factors that determine future CFC production. We are unsure about future
population levels, about the course of technological change, about new
products that will embody CFCs or substitutes, and so forth. Using a
combination of statistical techniques and informed judgment, we have
attempted to measure the extent of the uncertainty about future trends. To
take a particular case, we have surveyed experts on population and have from
them derived estimates of possible likely paths of population growth over
the coming decades. From these estimates, we have formed a judgment about
how likely different CFC production outcomes seem to be. When all of the
uncertain factors are incorporated, Table III-4 indicates, for example, that
it is equally likely that world CFC-12 (nonaerosol) production in 2020 will
lie between 490 thousand and 1600 thousand MT/yr as it is that it will lie
outside that range. That represents enormous uncertainty for a 35 year
glimpse into the future, but it is our view that such uncertainty accurately
reflects the current state of knowledge and ignorance about future CFC use
trend.
It should also be clear that relying upon maximum likelihood or best
guess scenarios to evaluate the severity of potential CFC damage could be a
serious mistake. If conditions in the future turn out to correspond with
the combinations of population growth, productivity growth, and CFC
-------
intensity that have here produced simulated trajectories above the best
guess, then the skewness or the results means that our present view of the
best guess will seriously underestimate the magnitude- of the future problem.
Table III-5 records an alternate measure of dispersion, the standard
deviations of the various important variables. Tracking these statistics
into the future reveals the ever increasing uncertainty that surrounds our
understanding. Computing the associated t-statistics, a dimensionless
composite measure of dispersion based on these standard deviations, supports
the validity of an interesting conjecture: emissions of nonaerosols, at any
point in time, should be less uncertain than the consumption that produces
them. As revealed in Table III-6, this conjecture is true, but only to a
limited degree. The estimated t-statistics for world consumption are,
indeed, always smaller than those for world emissions; i.e., the standard
deviation measured as a fraction of the corresponding mean is always higher
for consumption than for emissions. This result obtains, of course, because
depositing of nonaerosols into the bank delays their release into the
atmosphere, and thus spreads the full effect of their uncertainty into the
future. Still, the net. effect of the delay is spent in less than ten years.
-------
TABLE III-5
Estimated Standard Deviation!
Rased on Probabilistic Scenarios
1990
2000
2010
2020
2030
2040
2050
Major Driving Variables:
U.S. POPULATION (10**9)
ROW POPULATION (10**9)
US LABOR PRODUCTIVITY (1985$ 10**3/CAP*YR)
RW LABOR PRODUCTIVITY (1985$ 10**3/CAP«YR)
U.S GROSS OUTPUT (1985$ 10**12/YR)
ROW CROSS OUTPUT (1984i 10««12/YR)
WORLD CROSS OUTPUT (1985t 10"12/YJO
CFC-11 Nonaerosols:
RATIO OF CFC/CNP (OS) (10"3 KT/tt)
OS CFC CONSUMPTION (10««3 KT/TR)
ROW CFC CONSUMPTION U0«*3 MT/YR)
WORLD CFC CONSUMPTION (10"3 MT/tR)
WORLD CFC BANK (10««3 HI)
WORLD CFC EMISSIONS (10*»3 MT/YB.)
WORLD CFC CONCENTRATIONS (10**3 Ml)
CFC-12 Nouserosols:
RATIO OF CFC/CNP (US) (10*«3 MT/YR)
DS CFC CONSUMPTION (10**3 MT/YR)
ROW CFC CONSUMPTION (10"3 MT/YR)
WORLD CFC CONSUMPTION (10«*3 KT/YRJ
WORLD CFC BANK (10*«3 MT)
WORLD CFC HUSSION (10"3 MT/YR)
WORLD CFC CONCENTRAIONS (10«*3 XT)
CFC-11 Aerosols:
RATIO OF CFC/CNP (US) (10**3 MT/tt)
US CFC CONSUMPTION (10««3 KT/TR)
ROW CFC CONSUMPTION (10««3 MT/TR)
WORLD CFC CONSUMPTION (10"3 MT/YB.)
WORLD CFC BANK U0*«3 MT)
WORLD CFC EMISSIONS (10"3 MT/TR)
WORLD CFC CONCENTRATIONS (10««3 MT)
CFC-12 Aerosols:
RATIO OF CFC/CNP (US) (10**3 MT/YR)
US CFC CONSUMPTION (10*«3 MI/YR)
ROW CFC CONSUMPTION (10««3 MT/Ht)
WORLD CFC CONSUMPTION (10«*3 MT/YR)
WORLD CFC BANK (10»*3 MT)
WORLD CFC EMISSIONS (10**3 MT/YR)
0.000
0.119
1.377
0.193
0.336
1.049
1.091
3.642
18.599
45.190
57.042
221.859
32.897
193.966
7.359
37.292
43.100
71.635
280.414
41.200
241.392
4.455
1.141
41.262
41.831
0.000
41.832
199.079
5.836
1.455
64.706
65.565
0.000
65.565
0.004
0.280
3.537
0.476
0.922
3.049
3.166
7.327
50.918
138.286
167.643
606.510
115.892
845.606
15.242
104.597
133.179
209.456
763.267
145.321
1,039.144
9.199
3.204
129.989
131.388
0.000
131.388
1,002.714
11.997
4.060
200.137
202.298
0.000
202.298
0.007
0.493
7.227
0.882
1.976
6.425
6.682
10.589
99.508
330.046
372.635
1,353.727
274.083
2,6720.045
22.516
208.125
317.307
443.787
1,647.498
331.728
3,150.009
13.763
6.474
323.232
325.117
0.000
325.117
3,062.425
17.885
8.195
482.012
485.095
0.000
485.095
0.010
0.746
12.077
1.421
3.436
11.578
12.006
14.133
174.904
721.560
777.273
2,838.618
587.715
6,459.406
30.740
373.992
694.822
Ml. 720
3,291.197
676.156
7,459.504
18.958
11.731
736.073
738.389
0.000
738.389
7,844.797
24.572
14.861
1,069.223
1,073.188
0.000
1 .073.1RR
0.011
0.988
15.860
1.947
4.576
17.008
17.504
18.260
263.278
1,281.422
1,356.666
5,165.273
1,084.152
13.892.355
40.498
575.910
1,248.089
1,508.547
5,830.914
1,212.880
15,415.832
25.132
18.118
1,350.855
1,354.089
0.000
1,354.089
17,152.004
32.554
22.932
1,943.094
1,948.455
0.000
1 .04R.4SS
0.013
1.253
0.029
2.551
5.839
23.784
24.332
23.066
384.038
2,230.666
2,327.909
8,840.082
1,865.854
26,636.168
52.209
859.543
2,194.427
2,541.875
9,793.219
2,049.568
28,678.437
32.555
27.101
2,427.698
2,432.022
0.000
2,432.022
33,552.543
42.153
34.271
3,453.305
3,460.241
0.000
1.4*0.741
0.015
1.549
25.019
3.271
7.371
32.587
33.191
28.706
555.135
3,930.552
4,050.651
15,249.957
3.231.867
48,350.137
66.369
1,271.675
3,894.938
4,341.871
16,570.477
3,484.455
50,766.086
41.545
40.170
4.415.059
4,420.520
0.000
4,420.520
62,959.516
53.771
50.772
6.185.703
6.194.250
0.000
f, tQ/, .7 VI
-------
TABLE III-6
T-Statistics for Emissions and Consumption*
1990 2000 2010 2020 2030 2040 2050
CFC-11 Nonaerosols
World CFC Emissions
World CFC Consumption
CFC-12 Nonaerosols
World CFC Emissions
World CFC Consumption
7.06
4.82
7.22
4.87
2.96
2.42
2.95
2.41
1.86
1.62
1.91
1.68
1.31
1.18
1.41
1.28
1.01
0.92
1.11
1.03
2.81
0.75
0.91
0.85
0.66
0.61
0.76
0.71
* Computed as the ratio of the mean estimates of Table III-l and the
standard deviations of Table I1I-5.
-------
IV. Worldwide Regulation of Aerosol Consumption.
The results presented of the previous section do not reflect the
effects of any CFC regulation beyond the U.S. ban on nonessential use of
aerosols. They are advanced as benchmark projection against which the
ability of further regulation can be judged. They do not, therefore,
include the potential effects of either the voluntary cap on aerosol
capacity currently in place throughout the EEC or of other voluntary
restrictions on aerosol consumption now imposed by several other nations
around the world.
The present section will report results for aerosols under an
alternative assumption that all of the regulation presently in place
throughout the world, be it voluntary or mandated by law, will remain in
effect for the next seven decades regardless of economic circumstance.
Future economic forces may, of course, cause some countries to weaken their
restraint; serious future environmental damage may, on the other hand, drive
more countries to impose stringent constraints upon themselves. The future
regulatory environment is yet another source of uncertainty in projecting
ranges of emissions and concentrations for CFCs, but we will make no attempt
to endogenize the demand for regulation. A set of "worldwide status quo"
results is, instead, reported in this section to provide some insight into
the potential effectiveness of current worldwide efforts to control aerosol
emissions.
More specifically, we first assume that 1980 aerosol consumption in the
United States represents essential use. We apply, therefore, the frontier
growth model described in Sections I and II directly to those initial
levels. For the rest of the world, we assume (1) that the EEC capacity
-------
constraint will hold nonessential aerosol use at observed 1980 levels, (2)
that the EEC capacity constraint will allow essential aerosol use to grow
and (3) that voluntary restrictions imposed by individual, non-EEC nations
will hold their aerosol consumption fixed at 1980 levels (e.g., Japan: 25
percent reduction held fixed; Canada: 79 percent reduction held fixed;
Norway and Sweden: total ban held fixed). Finally, essential EEC use was
computed by applying the fraction of peak U.S. levels represented by 1980
U.S. consumption to the 1980 EEC consumption total.
The significance of these assumptions for the two aerosol compounds is
displayed in Table IV-1. Voluntary restraint across the world, if it were
to continue as is for the next 70 years, would effectively fix 73,400 metric
tons of CFC-11 aerosol consumption at 1980 levels. The remaining 38,500
tons of 1980 CFC-11 aerosol consumption (35 percent of total) would thus be
allowed to grow as a reflection of essential use in the U.S., essential use
in the EEC, or unrestrained essential and nonessential use across the rest
of the world. Similarly, existing regulation would fix 106,400 metric tons
of CFC-12 aerosol consumption leaving 72,400 tons (41 percent of 1980 world
consumption) to grow in one of the three categories just noted.
Table IV-2 reports the major results of incorporating these assumptions
into the probabilistic scenario analysis. All of the statistics recorded
there — the base cases, the mean trajectories, and the inner-quartile ranges —
\are all reduced by recognizing and projecting existing worldwide
regulation. Annual growth rates of total world consumption over the 70 year
time horizon fall, for the mean trajectory (e.g.), from 4.22 percent per
year to 2.86 percent per year for CFC-11 aerosols and from 4.20 percent per
year to 3.04 percent per year for CFC-12 aerosols.
-------
TABLE IV-1
Distribution of 1980 CFC Aerosol Consumption
Between Constrained and Unconstrained Components*
U.S. Essential
CFC-11 Aerosols:
Constrained
Uncons trained
4.100
EEC Essential
6.170
Rest of World
73.390
32.340
CFC-12 Aerosols:
Constrained
Uncons trained
5.400
10.730
106.400
61.670
*Measured in 10^ M tons of consumption.
-------
TABLE IV-2
Probabilistic Results—Consumption Responding to
Voluntary Regulation Across the Worlda
A. CFC-11 Aerosols
United States:
1980
2000
25th
Mean
75th
Median
4.10(nc)b
4.10(nc)
4.10(nc)
4.10(nc)
5.71 (nc)
8.11 (nc)
9.61 (nc)
7.17 (nc)
6.63
15.47
19.17
10.67
2020
(nc)
(nc)
(nc)
(nc)
2050
6.53
34.02
34.07
15.43
(nc)
(nc)
(nc)
(nc)
Rest of the World:
World:
B. CFC-12 Aerosols
United States:
Rest of the World:
World:
25th
Mean
75th
Median
25th
Mean
75th
Median
111.90(nc)
111.90(nc)
111.90(nc)
111.90(nc)
116.00(nc)
116.00(nc)
116.00(nc)
116.00(nc)
127.63 (157.61)
162.88 (260.07)
192.55 (346.31)
153.36 (229.49)
135.93 (166.61)
170.99 (268.18)
200.43 (355.55)
159.53 (236.66)
25th
Mean
75th
Median
25th
Mean
75th
Median
25th
Mean
75th
Median
5.40(nc)
5.40(nc)
5.40(nc)
5.40(nc)
178.80(nc)
178.80(nc)
178.80(nc)
178.80(nc)
184.20(nc)
184.20(nc)
184.20(nc)
184.20(nc)
7.29 (nc)
10.02 (nc)
12.28 (nc)
9.34 (nc)
217.43 (274.19)
278.97 (426.18)
329.81 (551.75)
254.17 (364.94)
227.17 (283.74)
288.99 (436.19)
339.16 (561.09)
263.51 (374.28)
140.62 (195.37)
307.41 (680.14)
418.92 (1,004.24)
219.41 (424.39)
155.34 (201.61)
322.88 (695.61)
429.59 (1,014.92)
230.09 (435.06)
8.14
17.98
22.64
13.90
(nc)
(nc)
(nc)
(nc)
260.43 (380.39)
553.79 (1,104.89)
716.12 (1,505.76)
376.07 (665.99)
275.16 (395.12)
571.77 (1,122.87)
719.43 (1,509.08)
389.97 (679.89)
141.63 (198.31)
826.79 (2,189.74)
939.68 (2,517.84)
313.17 (696.89)
168.54 (220.32)
860.81 (2,223.76)
993.32 (2,541.29)
328.60 (712.31)
8.42
38.08
40.74
20.09
(nc)
(nc)
(nc)
(nc)
283.31 (436.90)
1,504.28 (3,452.26)
1,666.93 (3,853.91)
548.89 (1,092.79)
318.87 (475.29)
1,542.36 (3,490.34)
1,752.06 (3,944.15)
568.99 (1,12.88
a Figures in parentheses are the comparable results in the absence of worldwide regulation (Section III).
b nc " no change.
-------
Finally, to test the potential effectiveness of unilateral U.S. aerosol
restriction, the model was manipulated to show that frontier growth rates
(CFC per GNP intensity) of -3.00 percent per year, -1.95 percent per year,
and -1.20 percent per year for the 1980-2000, 2000-2025, and 2025-2500 eras
would be sufficient to hold CFC-12 aerosol consumption constant along the
base trajectory for the entire 70 year time horizon. If these large
reductions in essential CFC use intensity were the result of economically
viable processes that were adopted by other producers around the world, then
the base case would lead to consumption in the rest of the world equal to
approximately 280,000 metric tons by the year 2050. World consumption
would, as a result, be 50 percent lower than the regulatory base case
recorded in Table IV-2, but still 50 percent higher than in 1980. Given
that the mean of the probabilistic runs is generally an order of magnitude
higher than the base case, even this extraordinary scenario does not allow
us to expect that world aerosol consumption will not increase markedly over
the next 70 years.
-------
Appendix A: Specification and Estimation of the Demand for CFCs.
The demand for CFCs is divided into two components. The first,
designed to depict experience in the United States, is driven by the CFC
frontier described in Section I. The second, designed to depict experience
outside the United States, is driven by U.S. demand and market development
in a way that is also described in Section I. The precise specifications
and supporting estimation procedures of these two modeling structures are
recorded in this appendix.
A-l. Specification of Demand in the U.S. - the CFC Frontier.
The fundamental construction underlying the U.S. model centers around
the specification of the potential frontier. An asymptotic CFC intensity,
denoted by Z*(t), is assumed to grow over time at a constant growth rate
plus a random error term reflecting changing technology, price trends, and
incomes. Accordingly,
(A-l) log [Z*(t)] - a + gt + e(t)
where
Z*(t) - potential CFC consumption per dollar of
real GNP in year t;
g — the growth rate of Z* (positive or negative);
a - an initializing parameter; and
e(t) - a random error term.
According to equation (A-l), the asymptotic frontier in Figure 1-2 should be
moving, on the average, at an annual rate of g. The estimate for this
-------
average g is thus taken to be the middle value for the frontier growth rate
in the scenario analysis. The high and low values are then simply the
reflection of the imprecision with which we can estimate that average g
given the uncertainty captured in (A-l) by the e(t).
The two stage estimation procedure employed in Stage I of this work
turned out to be too sensitive to the initialization assumptions, and has
been discarded. A second methodology based upon aggregate CFC proxies is,
instead, employed to estimate both the most likely value for g and its
surrounding range of uncertainty. More specifically, we consider, for each
of the four chemicals (CFC-11 and -12, aerosols and nonaerosols), a time
series of indexed proxies defined by
(A-2) CFC(t) - •£ a (1973)X.(t),
> J J
where
a (1973) - [CFC (1973)A1(1973)]
J J J
are the CFC input/output coefficients for products X computed using 1973
data. The a.(1973) are, in other words, the product- specific CFC-usage
intensities that were observed in 1973. Equation (A-2) uses these
intensities as weights in constructing a total use time series of derived
demand proxies based upon observed end product production levels recorded
over several decades. Estimates of income elasticities for the various CFC
chemicals can then be obtained by estimating the income elasticities of the
constructed CFC(t) proxies over recent years. The result is a set of
reliable estimates of frontier income elasticities as long as (1) future
patterns of use for the products are close to historical ones and (2) there
-------
is no systematic change in the CFC-usage intensities in future years. The
desired estimates of frontier growth rates can, finally, be produced by
subtracting 1 from the elasticity estimates.
The weights employed in this procedure are listed in Table A-l. They
are expressed in terms of percent of total CFC use in each product category
because the actual CFC(t) proxy series are index number series normalized
for 1973 - 100. Note, too, that CFC-11 and CFC-12 are distinguishable only
in nonaerosol form. The end product use time series for the 12 product
groups listed are drawn from the U.S. National Income and Product Accounts;
their precise sources are noted in Table A-2.
The resulting income elasticity estimates are recorded in Table A-3.
They conform, reasonably well, to less precise estimates that can be
obtained by computing the weighted average of end use income elasticities
produced by Houthakker and Taylor in 1970 (using, again, the weights of
Table A-l). The standard errors of the income elasticity estimates are,
however, too small to be applicable as reasonable measures of the
uncertainty with which we project Into the future.
As a way of estimating uncertainty about the growth of the CFC
frontier, we investigated the historical variation in growth rates of
different materials; Table A-4 displays the results. The first line simply
records the standard deviation of the growth rate of 13 major materials over
the period 1900 to 1984: a dispersion of 3.29 percent per annum. A second
measure is the same statistic for 24 resources (both fuel and nonfuel)
studied by the Paley Commission. Smaller dispersions are noted, probably
because the time period studied was one smaller structural change.
The most appropriate measure is the error of forecasts of 24 materials
-------
TABLE A-l
Weights for Estimating
the Derived Demand for CFCsa
New Construction
Household Furniture
Motor Vehicles
Textiles
Household Appliances
Marine Products
Electrical Equipment
Health Services
Food Products
Household Products
Personal Products
Industrial Products
Aerosols^
Nonaerosols
11
0
0
2.7
0
0
0
0
0
3.3
38.3
52.2
3.5
12
0
0
2.7
0
0
0
0
0
3.3
38.3
52.2
3.5
1J.
36.8
28.0
19.5
.9
9.2
2.2
0
0
3.4
0
0
0
12
19.8
56.9
3.9
0
4.0
.1
7.0
8.3
0
0
0
0
TOTAL
100.0
100.0
100.0
Notes:
a Expressed in percent of total in each category.
t> Source: Exhibit 2-1 (1973 data), RAND, 1986.
100.0
-------
TABLE A-2
Sources for Estimating
the Derived Demand for CFCs*
Source Series Name
New Construction Table 1.2, 1.12 NC
Household Furniture Table 2.5, 1.9 HF
Motor Vehicles Table 2.5, 1.4 MV
Textiles Table 2.5, 1.27 TX
Household Appliances Table 2.5, 1.10 HA
Marine Products Table 2.5, 1.82 MP
Electrical Equipment Table 1.2, 1.10 EE
Health Services Table 2.5, 1.71 HS
Food Products Table 2.5, 1.20 FP
Household Products Table 2.5, 1.37 CP
Personal Products Table 2.5, 1.67 PP
Industrial Products Table 1.4, 1.4 IP
*In each case, these refer to the U.S. National Income and Product
Accounts, Standard Tables, as In Survey of Current Business, July 1985.
Data are all in constant 1972 prices for the mentioned category.
-------
TABLE A-3
Results for Estimates of Derived Demand
for CFC's: 1948-84*
V _Q
Dependent Variable Output Elasticity Rho(p) R
Aerosols 1.02 .45 .999
(.065)
Nonaerosols-11 0.80 .40 .926
(.064)
Nonaerosols-12 1.09 .37 .979
(.041)
Notes:
a
Equation took the form ln(C. ) « A + bln(GNP + pu •, + e where
C " weighted production of CFC-containing products; GNP ™ real GNP
in 1972 prices; u and e " error terms;p m estimated coefficient
on first-order autoregressive error.
Figures in parentheses under output elasticity are the standard error of
the coefficient.
-------
TABLE A-4
Standard Deviation in Growth Rates
for Different Materials
Source
13 major materials, 1900-1984
Standard Deviation of
Growth Rate (percent per year)
3.29
Paley Commission (24 materials)
Actual 1950 vs. Actual 1972
Forecast vs. Actual 1972
2.49
1.94
Sources: Data for the major materials were taken from U.S. Department
of Commerce, Historical Statistics of the United States; data
on the Paley Commission, from Richard cooper and Robert
Lawrence, "The 1972-75 Commodity Boom," Brookings Papers on
Economic Activity, 1975.
-------
made by the Paley Commission for the period 1952 to 1972. It is
conceptually closest to the uncertainty about the forecast of CFC use
prepared for this study. We have rounded the figure to 2 percent per annum
when using it as the most likely forecast error for our CFC frontier growth
estimates.
A-2. Specification of Demand in the Rest of the World.
In order to estimate the growth of CFC consumption in the rest of the
world, it is necessary to estimate the demand for CFCs in regions outside of
the United States. This task poses formidable difficulties because of a
lack of data on consumption throughout the rest of the world, but
particularly outside of the OECD countries. Rough estimates are,
nonetheless, obtained in the following manner.
The basic assumption, as stated in the text, is that the trajectory of
consumption in the rest of the would will approach the U.S. experience as
per capita incomes outside of the U.S. approach American levels. To
represent this assumption formally, we consider
(A-3) InCCj^) - r In^/Y^ ,
where
C, - per capita consumption or production of CFCs in
region k;
k - 1 for the U.S., 2 for the EEC, 3 for developed
developed regions in the Pacific, 4 for the Eastern
Block, 5 for Latin America, 6 for Africa, and 7 for
all other countries;
Y, - per capita GNP for the same country groupings; and
-------
r — a coefficient to be estimated.
Data on CFC consumption, per capita incomes, and population are drawn from
the RAND report (1985). The CFC consumption data is recorded in Table B-13.
The unallocated totals, labeled "other used in developed countries," were
allocated proportionally to CFC production identified for each region. We
then estimated equation (A-3); the results are shown in Table A-5. Moving
from the estimates of r to most likely middle values for the desired
"relative intensity elasticities" recorded in Table II-2 of the text is
again accomplished by subtracting 1 from the values recorded for the four
chemicals in Table A-5. The ranges noted in II-2 are produced by assigning
values to the two outside 25 percent likelihood cells of plus and minus /2
times the standard deviations of the estimates also noted in Table A-5.
The results confirm, in general, the initial impression that other
countries lag far behind the U.S. in their use of CFCs. We may, therefore,
witness very rapid growth in CFC production and consumption in the rest of
the world if, in the absence of regulatory restraint, economic growth
continues at historical levels in those countries.
-------
Table A-5
Estimates of Cross Section Regression
—Per Capita Income Elasticities
Chemical Elasticity (Standard Error of Coefficient) R2
11 Nonaerosols 2.27 (.25) .848
12 Nonaerosols 2.15 (.20) .868
11 Aerosols 2.30 (.37) .729
12 Aerosols 1.93 (.27) .806
-------
Appendix B: The Data
This appendix begins with a review of the procedure used to construct
the distributions for the exogenous variables identified in Section I of the
report: the rate of growth of population and productivity in the U.S. and
the rest of the world from 1980 through the year 2050. Each distribution is
constructed to reflect not only statistical dispersion around published
estimates, but also judgmental dispersion evidenced by disagreement across
experts' best guesses. A brief description of the CFG data used to
estimate the demand structure is also provided.
A. Population.
The population growth estimates for the U.S. and the rest of the world
are both computed from a distribution of global growth rates deduced from
the estimates recorded in Table B-l. The upper part of the table records
the projected rates of growth of the world population reported in 9 separate
studies published since 1980. The averages of these cited estimates are
caken to be the maximum likelihood estimates for world population growth for
the periods noted. For the years 1980 through 2000, for example, the
published studies include estimates ranging from 1.3 percent per year up to
1.7 percent per year with an average value of 1.6 percent. The middle cell
of a three cell distribution approximating the true distribution of our
uncertainty about future population growth rates should therefore be
assigned the value 1.6 percent.
For this study, however, we needed more than point estimates of the
most likely growth estimates; we also needed measures of the uncertainty
associated with each future projection. One possible source of these
measure was the cited authors' own estimates of uncertainty, but few authors
1
-------
Table B-l
Population Growth Estimates - Worlda
N-Y (1983)
EPA (1985)
UN (1985)
World Bank (1984)
Population Reference
Bureau (1985)
IIASA (1981)
Rotty and
Marland (1980)
OECD (1981)
Keyfitz (1985)
mean
(disagreement
dispersion)
1980-
2000
1.7
(.2)
1.6
1.6
(.13)
1.6
1.6
1.7
1.3
1.6
1.4
(.3)
1.6%
(.13%)
2000-
2025
1.1
(.36)
0.9
1.3
(.41)
1.2
1.2
0.9
1.3
-
0.9
(.3)
1.1%
(.182)
2025-
2050
0.3
(.35)
0.4
0.7
(.29)
0.7
_
-
_
-
0.4
(.3)
0.5%
(.15%)
judgemental dispersion
(mean) .21% .36% .32%
Note:
a The numbers in parentheses represent cited or deduced judgemental
estimates of standard deviation contained in source.
-------
provide these estimates. Indeed, only 3 authors cited in Table B-l provided
error bounds of any kind for their forecasts. For these three estimates,
though, the average of the judgmental uncertainty ranges does provide one
measure of dispersion. It is identified as judgmental dispersion in Table
B-l and elsewhere; for the period from 1980 through 2000, an average of .21
percent per year is obtained.
A second method of estimating the uncertainty around future trends
involves looking at the extent to which experts disagree in their
projections -- this measure is called the disagreement dispersion.
According to this technique, we examine, in the case of population covered
in Table B-l for example, disagreement across the 9 published estimates.
Taking these estimates as data and estimating the extent of uncertainty by
computing the standard deviation, we find a value of .13 percent per year.
This second estimate is of the same order of magnitude as the judgmental
dispersion just noted.
Studies by behavioral psychologists and others have noted a nearly
universal tendency of people to underestimate the uncertainty associated
with many events. Even experts with knowledge of statistics display this
tendency. In a preliminary paper designed to quantify this tendency,
Nordhaus has considered the issue of translating disagreement across
projections of the same event into a measure of the true uncertainty of the
event. His results, rough but highly suggestive, imply that both measures
of dispersion need to be increased by a factor of 50 percent before they can
be considered reasonable indicators of the true uncertainty surrounding our
knowledge about future forecast events such as population growth rates over
the next 15 years. Table B-2 records this expansion in columns (D) and (F)
-------
Table B-2
Population Growth Estimates - World a
mean disagreement
years estimate dispersion
(A)
1980-2000
2000-2025
2025-2050
(B)
1.6%
1.1%
0.5%
(C)
.13%
.18%
.15%
adjusted
disagreement
(D)b
.20%
.27%
.23%
Judgemental
dispersion
(E)
.21%
.36%
.30%
adjusted
judgemental
(F)c
.32%
.54%
.45%
deduced
dispersion
(G)d
.30%
.40%
.35%
one standard
deviation range
(H)
1.3%/1.9%
o.2%/o!e%
Notes:
a Sources noted in Table B-l; based on 9, 8, and 5 published
estimates for the three noted time periods respectively.
b Computed as 1.5 times column (C).
c Computed as 1.5 times column (E).
d Deduced estimation of standard deviation judged from the values
noted in columns (D) and (F) adjusted in later eras to maintain
consistency with the larger sample estimates of earlier eras. The
associated t-stats decline as the estimates apply further into the
future (5.33; 2.75; 1.42).
-------
and ultimately proposes a deduced dispersion estimate for one standard
deviation around the sample mean of .3 percent. The final assertion,
representing a slight adjustment for the sake of consistency within and
across time periods in the future, produces a one standard deviation range
of possible population growth rates of 1.3 percent per year and 1.9 percent
per year around a mean estimate of 1.6 percent. The very same process
produces the recorded ranges for the other two time periods.
World population growth estimates are not, of course, estimates for the
United States. To produce ranges of U.S. estimates, the well accepted
Keyfitz projections of .68 percent, .42 percent and .08 percent per year are
employed for the three time periods, respectively. Requiring that the
ranges around these estimates have the same t-statistics as the world
estimates recorded in Table B-2, the estimates recorded for the U.S. in
Table B-3 are easily obtained. Notice, though, that the three possible
values are labeled "high", "medium" and "low" instead of plus and minus one
standard deviation. That is because the ranges recorded in Table B-3
represent the mean plus and minus the square root of 2 times the deduced
dispersion estimate of one standard deviation. Why? Because a mean and a
standard deviation fully specify a normal distribution for a random
variable; and if that distribution is to be approximated by a three cell
discrete distribution with 25 percent, 50 percent and 25 percent probability
weights, then the values assigned to the two outside, 25 percent cells must
be the mean plus and minus /2 times the standard deviation.
The rest of the world estimates recorded in Table B-3 were produced
similarly, but have been adjusted slightly from the world estimates by a
weighting process that makes them consistent with (1) the mean U.S.
-------
Table B-3
Population Growth Ranges8
United States (percentage change per year)
Era 1980-2000 2000-25 2025-50
High 0.8 0.6 0.15
Middle 0.7 0.4 0.1
Low 0.6 0.3 0.0
Rest of World (percentage change per year)
Era 1980-2000 2000-25 2025-50
High 2.0 1.5 0.9
Middle 1.7 1.1 0.5
Low 1.3 0.7 0.2
a Source: Table B-2; see text for procedure.
-------
estimates based on Keyfltz and reported in Table B-3, (2) the mean world
estimates of Table B-2, (3) the standard deviations supporting the U.S.
ranges of Table B-3, and (4) the standard deviations reported in Table B-3.
The procedure that assures these consistencies is entirely algebraic.
Given, for example, an aggregate random variable, Z, that is the weighted
average of two other random variables, X and Y, according to
Z - aX + (l-a)Y,
knowledge about the distribution of Z and X can be translated into knowledge
about the distribution of Y according to
mean(Y) - {[mean(Z)]-a[mean(X)]}/(l-a), and
var(Y) - {[var(Z)-a2[var(X)]}/(l-a)2,
assuming that X and Y are uncorrelated. Knowledge about the distribution
of population growth estimates for the world and the U.S. is thus sufficient
to produce consistent estimates about the distribution of population growth
across the rest of the world.
. Productivity Growth.
Productivity growth estimates for the U.S. and the rest of the world
are based upon a similar procedure. There exists a collection of studies
that report real GNP forecasts for the United States. A summary of their
content is found on Table B-4, and its import for present purposes indicated
in Table B-5. Assuming an independence between population growth and GNP
growth that might not be justified, the estimates and ranges in Table B-5
for GNP and Table B-3 for population are combined in Table B-6 to produce
productivity growth estimates and ranges for the U.S. Tables B-7 and B-8
display the results of repeating the process for world productivity growth
projections so that Table B-9 records estimates and ranges for the rest of
4
-------
Table B-4
GNP Growth Estimates - U.S.
Exxon (1980)
USDOE (1980)
H-K (1979)
RFF (high) (1980)
RFF (low) (1980)
WEC (high ) (1978)
WEC (low) (1978)
IIASA (high) (1980)
IIASA (low) (1980)
Mean
Judgmental
dispersion
1980- 2000-
2000 2025
2.7
2.5
3.0
2.7
2.3
3.5 3.1
2.5 2.5
3.7 2.2
2.4 1.0
2.89 2.20
.49 .88
2025-
2050
-
-
-
-
-
-
-
-
na
na
-------
Table B-5
GNP Growth Estimates—U.S.a
mean judgmental adjusted deduced one standard
Years estimate dispersion judgmental dispersion deviation range
(A) (B) (C) (D)b (E)c (F)
1980-2000 2.81% .49% .74% .7% 2.1%/3.5%
2000-2025 2.20% .88% 1.32% 1.3% 0.9%/3.5%
2025-2050 Data not available; world data used for both U.S. and rest of world.
a Source: Table B-4 based on 9 and 4 published estimates.
b Computed at 1.5 times column (C).
c Deducted estimation based on column (D). The associated t-stats as the estimates apply
further into the future (3.87 and 1.69).
-------
High
Middle
Low
Table B-6
Productivity Growth Rates - U.S.a
1980-2000
3.2
2.1
1.0
2000-2025
3.3
1.6
0.0
2025-2050b
2.2
1.1
0.0
a Source: Tables B-5 and B-3 assuming independence between
population growth and GNP.
° The world figures are used for both the U.S. and the rest of the
world because of data deficiencies.
-------
Table B-7
Productivity Growth Estimates - Worlds
N-Y (1983)
OECD (1981)
IEW (1985)
IIASA (1981)
Kahn (1977)
Lovins (1981)
E/R (1985)
Rotty and
Marland (1980)
RTF (1980)
EPA (1985)
WEL (1977)
Hudson
mean
(disagreement
dispersion)
judgemental
dispersion (mean)
1980-
2000
2.3
(0.7)
2.7
(.62)
1.47
(.67)
1.75
(.78)
2.75
1.29
2.67
(.67)
1.49
2.0
(.57)
1.60
1.74
(.76)
2.8
2. OX
(.59%)
.682
2000-
2025
1.6
(0.5)
-
2.77
(.47)
1.4
(.71)
3.1
0.9
-
1.50
1.95
(.21)
1.68
1.73
(.76)
1.4
1.8%
(.701)
.54Z
2025-
2050
1.0
-
-
1.19
0.5
—
-
-
1.67
-
.98
(.32)
1.1X
(.49%)
.32Z
Notes:
a The numbers in parentheses represent cited or deduced judgemental
estimates of standard deviation contained in source.
-------
Table B-8
Productivity Growth Estimates - World3
mean disagreement adjusted judgmental
years estimate dispersion disagreement dispersion
(A) (B) (C) (D)b (E)
adjusted
judgmental
deduced
dispersion
(G)d
one standard
deviation range
(H)
1980-2000
2000-2025
2025-2050
2.0%
1.8%
1.1%
.60%
.70%
.49%
.90%
1.35%
.74%
.68%
.54%
.32%
1.02%
.81%
.48%
.95%
1.00%
.60%
1.05%/2.95%
.80%/2.80%
.50%/1.70%
NOTES:
a Sources noted in Table B-7; based on 12, 10, and 5 published estimates for the three noted time
periods respectively.
b Computed as 1.5 times column (C).
c Computed as 1.5 times column (E).
d Deduced estimation of standard deviation judged from the values noted in columns (D) and (F) adjusted
in later eras to maintain consistency with the larger sample estimates of earlier eras. The
associated t-stats decline as the estimates apply further into the future (2.11; 1.71; 1.38).
-------
Table B-9
Productivity Growth Ranges - Rest of World3
Era
High
Middle
Low
1980-2000
2.9
2.0
1.1
2000-2025
2.8
1.8
0.9
2025-2050
1.7
1.1
0.5
a Source: Tables B-8 and B-6; see text.
-------
the world that are consistent with both the world estimates of Table B-8 and
the U.S. estimates of Table B-6. The ranges reported in Tables B-6 and B-9
are not expanded by the same multiplicative factor of /2 to preserve both
the mean and the variance of the underlying normal distribution. The
standard deviation estimates are used, instead, to reflect the effect of the
likely negative correlation between population growth and productivity. The
procedure that assures consistency in moving not only from the U.S. and the
world estimates to the rest of the world, but also from growth in total GNP
to growth in per capita GNP, is identical in spirit if not exact detail to
the algebraic process described above.
C. CFC Data.
Two types of CFC emissions data are required to estimate the demand
structure. The first, a time series of U.S. production for CFC-11 aerosols,
CFC-12 aerosols, CFC-11 nonaerosols, and CFC-12 nonaerosols, was found in
the TCF study. They are recorded here in Tables B-10 through B-12. It is
important to note that these series were, to a significant: degree,
constructed on the basis of a wide range of assumptions recorded in the
attached notes. They may be the best data available, but they are therefore
far from perfect from a statistical perspective. They are also production
data and not the consumption data that would usually be employed in demand
analysis.
A second set of data, cross sectional production data for a specific
year, was also required (although, again, consumption data would have been
more appropriate). Table B-13 records the set that has been employed. The
text surrounding its presentation as Table 4.1 in the January 1986 RAND
study [Quinn, et. al. 1986] outlined a series of steps employed to construct
-------
Table B-10
U.S. CFC Production - Total^
CFC-llb CFC-12b
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
Notes:
a Source: IGF Table A-l.
b U.S. ITC. Synthetic Organic Chemicals. Annual
Series. These include CFC-22 used as an intermediate.
Approximately 28 percent of CFF-22 is used in the
production of teflon and other products.
22.90
27.40
32.80
41.20
56.60
63.60
67.40
77.30
77.30
82.70
92.70
108.20
110.90
117.00
135.90
151.40
154.70
122.30
116.20
96.40
87.90
75.80
71.70
73.80
63.70
73.10
59.60
71.30
75.50
78.70
94.30
98.60
103.40
123.10
129.90
140.50
147.70
166.80
170.30
176.70
199.20
221.70
221.10
178.30
178.30
162.50
148.40
133.30
133.80
147.60
117.00
134.30
-------
Table B-ll
U.S. CFC Production - Aerosals*
Year CFC-11 GFC-12
1958 19.90t> 35.80C
1959 23 10 42.80
I960 26.70 45.30
1961 32.40 47.20
1962 43.20 56.60
1963 46.60 59.10
1964 47.50 62.00
1965 52.40 73.90
1966 50.20 77.90
1967 51.50 84.30
1968 55.20 88.60
1969 61.40 100.00
1970 60.004 102.10d
1971 63.10 106.20
1972 72.10 117.50
1973 83.40 135.20
1974 86.70 123.90
1975 73.30 94.50
1976 64.00 80.30
1977 36.50 47.30
1978 32.906 41.906
1979 9.50 5.30
1980 9.50 5.80
1981 9.50 5.60
Notes:
a Source: ICF Table A-2
b Estimates for 1958-1969: based on the assumption that in 1958 the
U.S. dominates the OECD market as CFC-11. In 1958, the CMA data
(CMA. Production, Sales, and Calculated Release of CFC-11 and 12
through 1982; Expanded Data) show a market share of open cell.
aerosol, and all other production to be 87 percent of total CFC-11
production. Thus, the U.S. aerosol share of total CFC production is
assumed to be 87 percent in 1958, declining smoothly to 54 percent
in 1970, where 54 percent is the aerosol proportion in 1970
described in footnote C. Total CFC-11 production is bsed on U.S.
ITC data.
c Estimates for 1958-1969: Based on the assumption that CFC-12
aerosol production was 60 percent of total CFC-12 production, as it
was in 1970.
d Data for 1970-1977: U.S. EPA. Regulation Chlorofluorocarbon
Emissions; Effects on Chemical Production.(EPA-560/12-80-0016)
October 1980, pp. 73, 74, 80. Reports production figures from pp.
73-74, using aerosol/nonaerosol proportions from p. 80.
e Estimates for 1978 and 1981 computed by subtracting usage in
non-aerosol applications from total CFC usage. See Tables 10 and 12
-------
Table B-12
U.S. CFG Production - Nonaerosalsa
Year CFC-11 CFC-12
1958 OQ 23.8QC
1959 A. 30 28.50
1960 6.10 30.20
1961 8.80 31.50
1962 13.40 37.70
1963 17.00 39.50
1964 19.90 41.40
1965 24.90 49.20
1966 27.10 52.00
1967 31.20 56.20
1968 37.50 59.10
1969 46.80 66.80
1970 50.80d 68.00d
1971 54.00 70.80
1972 64.00 81.70
1973 68.00 86.60
1974 68.00 97.10
1975 49.00 83.90
1976 52.20 98.00
1977 59.90 115.20
1978 55. OO6 106. 50e
1979 66.30 128.00
1980 62.20 128.00
1981 64.30 142.00
1982 63.70f 117. 00f
1983 73.10 134.30
Notes:
a Source: ICF Table A-3
b Estimates for 1958-1969: CFC-aa non-aerosols are defined as CFC-11
total minus CFC-11 aerosol (see table A-2).
c Estimates for 1958-1969: based on the assumption that CFC-12 nonaerosol
production was 40 percent of total CFC-12 production, as it was in 1970.
d Data for 1970-1977: U.S. EPA. Regulation Chlorofluorocarbon
Emissions; Effects on Chemical Production. (EPA-560/12-60-0016)
October 1980, pp. 73, 74, 80. Reports production figures from pp.
73-74, using aerosol/nonaerosol proportions from p. 80.
e Estimates for 1978-1981: OECD. Economic Aspects of CFC Emission
Scenario*? prepared by the Chemical Groups and Management Committee, 15
September 1982, p. 32. Based on the assumption that the split between
CFC-11 and cFC-12 in these years was the same as it was in 1977.
Estimates for 1982-1982: based on the assumption that production for
aerosol applications is zero, which makes nonaerosol production equal to
total production.
-------
Table B-13
1980 Worldwide CFC Production3
CFC-11 CFC-12
Region Aerosol Nonaerosol Aerosol Nonaerosol
Developing Nations 102.1 181.9 131.8 207.3
U.S. 4.1 67.6 5.4 128.4
European Community 56.4 57.6 74.1 33.9
Pacific 8.2 35.7 10.8 36.9
Other Uses in Developed
Nations 33.4 21.0 42.6 8.1
East Bloc 10.7 8.8 45.2 36.9
Developing Nations 3.2 2.6 6.2 5.0
Latin America 1.7 1.4 3.4 2.7
Africa 1.4 1.1 2.2 1.8
Other 0.1 0.1 0.6 0.5
Total 116.0 193.3 184.2 249.2
GRAND TOTAL 309.3 433.4
a Source: Rand Table 4.1.
(8304b)
-------
even this one year cross section from a variety of sources. Of particular
note is the row labeled "Other Uses in Developed Nations" into which many
apparent discrepancies were collected. Production listed there was, for the
purposes of estimation in this study, assumed to be distributed
proportionately across the non-U.S. developed countries.
-------
Appendix C: The Complete Model.
The complete set of equations for the CFC projections produced in this
report is recorded here. The form of the equations is identical for all
four chemicals, but the parameters differ according to the statistical
estimates.
Equations (a) through (d) represent the exogenous projections of
populations (P) and per capita GNP (A) in the United States (US) and the
rest of the world (ROW):
(a) PUS(t) exogenous {3 paths):
(b) PROW(t) exogenous {3 paths);
(c) AUS(t) exogenous {3 paths); and
(d) AROW(t) exogenous {3 paths).
Equation (e) represents the growth in the CFC frontier in the United States,
where Z* is potential use of CFCs per unit GNP.
(e) log [Z*(t)] - a + gt. {3 frontier paths)
The ratio of actual to potential CFC consumption in the U.S. [Z/Z*] is given
by a logistics curve for each frontier:
(f) W(t) - -b-ct {1 for each frontier)
where the definition of W(t) is
(g) Z(t) - Z*(t)/U + exp[W(t)J) (identity)
-------
Total CFG consumption in the U.S. is given by:
(h) CFCUS(t) - Z(t) PUS(t) AUS(t) {identity}
Total CFC consumption in the rest of the world is:
AROW(t) E
(i) CFCROW(t) - Z(t) AROW(t) PROW(t)
AUS(t)
with 3 values of E. Finally, the emissions of CFCs into the atmosphere,
ECFC(t), and the volume of the banks, BCFC(t), are be determined by Koyck
lag equations of the form
(j) ECFC(t) - (1-d) BCFC(t-l) + .1 CFCN(t) and
{3 sets of equations defined by 3 d parameters)
(k) BCFC(t) - d BCFC(t-l) + .9 CFCN(t)
for nonaerosols and
(j') ECFC(t) - CFCA(t) and
{1 set of equations)
(k') BCFC(t) - 0
-------
References
Camm, Frank, and James K. Hammitt, An Analytic Method for
Constructing Scenarios from a Subjective Joint Probability
Distribution, The Rand Corporation, N-2442-EPA, May 1986.
Chemical Manufacturers Association, Fluorocarbon Program Panel,
World Production and Sales of Chlorofluorocarbons CFG-11 and
CFG-12, various years.
Exxon Corporation, World Energy Outlook, New York, 1980.
Gibbs, Michael, ICF, "Scenarios of CFC Use: 1985-2075," presented
at "Protecting the Ozone Layer: Workshop on Demand and
Control Technologies," U.S. EPA, Washington, D.C., March 6,
1986.
Gibbs, Michael, ICF, "Summary of Historical Chlorofluorocarbon
Production," ICF, presented at "Protecting the Ozone Layer:
Workshop on Demand and Control Technologies," U.S. EPA,
Washington, D.C., March 6, 1986.
Hafele, W., Energy In a Finite World, IIASA, Ballinger Publishing
Co., Cambridge, Massachusetts, 1981.
Keyfitz, Nathan, et al., "Global Population 91975-2075) and Labor
Force (1975-2050)," Institute for Energy Analysis, Oak Ridge
Associate Universities, Oak Ridge, Tennessee, 1983.
Lovins, A.B., L.H. Lovins, F. Crause, and W. Bach, Energy Strategies
for Low Climate Risks, prepared for the German Federal
Environmental Agency, San Francisco International Project for
Soft-Energy Paths, June 1981.
My T. Vu, World Population Projections 1984, World Bank, Washington,
D.C., 1984.
National Academy of Sciences, Causes and Effects of Changes in
Stratospheric Ozone; Update 1983, Washington, D.C., 1984.
National Academy of Sciences, Causes and Effects in Stratospheric
Ozone; An Update, Washington, D.C., 1982.
National Academy of Sciences, Halocarbons: Effects on Stratospheric
Ozone, Washington, D.C., 1976.
National Academy of Sciences, Protection against Depletion of
Stratospheric Ozone, Washington, D.c7, 1979.
Nordhaus, W.D. and G. Tohe, "Future Carbon Dioxide Emissions From
Fossil Fuels," Changin Climate, National Academy of Sciences,
Washington, D.C., 1983.
-------
References (Continued) -2-
Palmer, Adele R., William E. Mooz, Timothy H. Quinn, and Kathleen A.
Wolf, Economic Implications of Regulating Chlorofluorocarbon
Emissions from Nonaerosol Applications, The Rand Corporation,
R-2524-EPA, June 1980.
Population Reference Bureau, 1985 World Population Data Sheet,
Washington, D.C., 1985.
Quinn, Timothy, Kathleen A. Wolf, William E. Mooze, James K.
Hammitte, Thomas W. Chesnutt, and Syam Sarma, Projected Use,
Emissions, and Banks of Potential Ozone Depleting Substances,
The Rand Corporation, N-2282-EPA, January 1986.
Reilly, John, Rayola Dougher, and Jae Edmonds, Determinants of
Global Energy Supply to the Year 2050, Institute for Energy
Analysis, Washington, D.C. 1981.
Ridker, R.G. and W.D. Watson, To Choose a Future, Baltimore,
Maryland, The Johns Hopkins University Press, 1980.
Rotty, R.M. and G. Marland, Constraints on Carbon Dioxide Production
From Fossil Fuel Use, ORAU/IEA-80-9(M), Institute for Energy
Analysis, 1980.
Seidel, Stephen, and Dale Keyes, Can We Delay a Greenhouse Warming?,
U.S. Environmental Protection Agency, EPA-230-10-84-001,
Washington, D.C., 1983.
Tversky, Amos and Daniel Kahneman, "Judgment Under Uncertainty:
Heuristics and Biases," Science, September 1974.
U.S. Department of Energy, Annual Report to Congress,
DOE/EIA-0173C80), U.S. Department of Energy, Washington, D.C.,
1980.
United Nations, World Population Prospects; Estimates and
Projections as Assessed in 1982, New York, New York, 1985.
Wolf, Kathleen A., Regulating Chlorofluorocarbon Emissions; Effects
on Chemical Production, The Rand Corporation, N-1483-EPA,
August 1980.
World Energy Conference (WEC), The Full Report to the Conservation
Commission of the World Energy Conference: World Energy
Demand, New York, IPC Science and Technology Press, 1981.
(1014c)
-------
SCENARIOS OF CFC USE: 1985-2075
by
Michael J. Gibbs
Prepared for
Strategic Studies Staff
Office of Policy Analysis
Office of Policy, Planning and Evaluation
U.S. Environmental Protection Agency
401 M Street, S.W.
Washington, D.C. 20460
February 1986
-------
EXECUTIVE SUMMARY
Using recent historical trends as a guide, potential future use and
emissions of chlorofluorocarbons (CFCs) are projected in a series of
scenarios. The total use of CFCs 11, 12, 22, and 113 worldwide is likely to
increase between 1985 and 2075 at an annual rate of approximately 2.4 percent
to 3.5 percent per year. Use per capita is likely to increase at a rate of
1.7 percent to 2.7 percent per year. At these potential growth rates, total
use of these CFCs may grow from the current annual level of approximately 1.2
million metric tons to between 10.0 and 27.5 million metric tons by 2075.
The size of this range is driven by the uncertainties regarding the
primary determinants of future CFC consumption: population growth and
economic growth. A range of projections of global and regional population and
Gross National Product (GNP) was used to develop the CFC use scenarios.
Additionally, assumptions about the roles of CFCs in future economies were
varied.
-------
TABLE OF CONTENTS
Page
EXECUTIVE SUMMARY 1
1. INTRODUCTION 2
2. DATA AND METHODS 3
2.1 Data Sources for CFC Use and Emissions 4
2.2 Statistical Analysis 5
2.3 Implementation of the Scenarios 7
3. RESULTS 14
4. LIMITATIONS 18
4.1 Limitations in Data 18
4.2 Limitations in Methods 19
4.3 Next Steps 20
ANNEX A A-1
ANNEX B B-l
ANNEX C C-l
ANNEX D D-l
ANNEX E E-l
-------
-2-
I. INTRODUCTION
This paper presents scenarios of the potential future use of
chlorofluorocarbons (CFCs). CFCs are a non-toxic, non-flammable class of
chemicals used as propellants in aerosol cans and for a variety of non-aerosol
applications (such as in making plastic foams and as the coolant in
refrigerators, air conditioners, and freezers). In 1974, it was hypothesized
that CFCs emitted to the atmosphere will reduce the amount of ozone in the
stratosphere,1 thereby allowing greater amounts of ultraviolet radiation to
reach the earth's surface. Since that time, considerable experimentation,
data collection, atmospheric monitoring, and modeling have been used to
investigate the potential for CFCs to induce stratospheric ozone depletion,
and the implications of such depletion. In response to concerns over ozone
depletion, public use of CFCs as an aerosol propellant decreased in the United
States and abroad between 1974 and 1978 by two-thirds. In 1978 the use of
CFCs in nonessential aerosol propellant applications was banned in the United
States. Norway, Sweden, and Canada have adopted similar restrictions and the
European Economic Community has adopted a 30 percent cutback in the use of
CFCs as aerosol propellants.
Due to concerns about potential ozone depletion that may be caused by CFC
emissions, the United States and other countries have initiated multilateral
negotiations regarding future uses and restrictions of CFCs. Cooperation
among all countries is required in order to control the effects of CFC
emissions because the impacts of concern are global in nature; i.e., the
emissions have the same effects on stratospheric ozone regardless of their
point of origin. In support of these negotiations, this analysis presents
potential future scenarios of CFC use and emissions worldwide through the year
2075.
The effects of today's CFC use will be felt for a long time because the
chlorine released from CFCs remains in the stratosphere for a very long time,
up to hundreds of years. If the chlorine that today's CFCs put into the
stratosphere cause problems in the future (e.g., by causing ozone depletion
after 2000), the only method we will have to correct those problems will be to
reduce CFC use at that time, i.e., in the future. We will be unable to
remove from the stratosphere the chlorine put there by today's uses of CFCs.
Therefore, to evaluate the potential implications of policies to reduce CFC
use now or in the future, scenarios of CFC use into the fairly far off future
are required.
1 M.J. Molina and F.S. Rowland, "Stratospheric Sink for
Chlorofluoromethanes: Chlorine Atom Catalysed Destruction of Ozone," Nature,
1974, pp. 810-812.
-------
-3-
This need to have long range scenarios poses a challenge to policy
analysts. Most factors that will influence the use of CFCs after 2000 can be
projected only with great uncertainty. Describing the potential changes in
the economies and populations of developed and developing countries beyond
2000 is inherently uncertain. Adding to this demographic and economic
uncertainty is uncertainty regarding the roles that CFCs will continue to play
in the economies of various countries around the world. Within the next 90
years unforeseen technologies may develop that replace today's major uses of
CFCs or require additional use of CFCs. The approaches developed here reflect
the key uncertainties in the analysis. All the assumptions that influence the
scenarios are presented, and a range of possible scenarios is described.
Even with the most advanced methods, however, scenarios into the far-off
future must be considered somewhat speculative. The later years of the
scenarios can only identify what could possibly occur, but cannot say with
any definitiveness what will actually occur.
This paper first describes in section 2 the approach used to develop the
scenarios. Section 3 presents results of the analysis and section 4
summarizes the limitations of the scenarios.
2. DATA AND METHODS
This section summarizes the data and methods used to project scenarios of
CFC use and emissions worldwide. The general approach is in three parts:
• collect data on the historical use of CFCs;
• analyze the data statistically to identify
relationships between CFC use and population, GNP per
capita, and the prices of CFCs; and
• use the results of the statistical analysis to project
future use and emissions, also using assumptions about
future population and economic growth.
First the availability of historical CFC use data is described.
Then the statistical analysis is summarized. Finally, the manner in
which the results of the statistical analysis are used to generate
the scenarios of CFC use and emissions is presented.
-------
-4-
2.1 Data Sources for CFC Use and Emissions
Data describing historical CFC use and emissions were collected from U.S.
Environmental Protection Agency (EPA) reports, U.S. International Trade
Commission (ITC) publications, Chemical Manufacturers Association (CMA)
reports, and Organisation for Economic Cooperation and Development (OECD)
reports. A list of the data used and their sources are presented in Annex A.
Data were generally available for the use of:
• CFC-11, CFC-12, and CFC-22 in the U.S. from 1958 to
1982;
• CFC-113 in the United States from 1958 to 1979;
• CFC-11, and CFC-12 in the OECD countries from 1958 to
1982; and
• CFC-22 in the OECD countries from 1958 to 1975.
However, the division of total CFC use for CFC-11 and CFC-12 between aerosol
and non-aerosol applications was not available for the U.S. between 1958 and
1969 and not available for the other OECD countries between 1958 and 1975.
The distinction between aerosol and non-aerosol uses is important in order to
project separate scenarios of future aerosol and non-aerosol uses. The
separate scenarios are needed because different control policies exist for the
two types of applications (and because different future control policies may
be contemplated). The assumptions made regarding the aerosol/non-aerosol
division are described in Annex A.
Data were generally not available or not reliable for the following:
• CFC-113 in the OECD countries other than the U.S.; and
CFC-11, CFC-12, CFC-22 and CFC-113 in countries
outside the OECD.
This lack of data is particularly troublesome because: (1) the use of CFC-113
is expected to grow significantly (because it is used in the production of
electronic components); and (2) the countries outside the OECD have the
overwhelming majority of the world's population, and their population and per
capita GNP are projected to grow at a faster rate than in the developed
world. Consequently these countries may represent potentially large future
users of CFCs. As described below, alternative approaches (that are not based
on statistical analyses of historical data) were used to project scenarios of
potential use for these CFCs and portions of the world for which historical
data were not available.
-------
•o-
Data were also collected on the rate of emissions of CFCs. Emissions do
not equal use for some CFC applications because the CFCs remain in some
products for many years (e.g., refrigerators). Annex C describes the data
collected on emissions rates. The emissions rates were used along with the
scenarios of future use to develop scenarios of future emissions.
2.2 Statistical Analysis
The objective of the statistical analysis was to describe the historical
behavior of CFC use over time using common economic and demographic variables
such as population and Gross National Product (GNP). These variables were
used because CFC use can be shown to be related to them and because there are
a variety of projections for them that stretch far into the future.
Basing the analysis on population and GNP is appropriate because the use
of most products and commodities in our modern economies is generally related
to the size of the economy and its wealth, or standard of living. For
example, in various developed countries the use of products such as cars,
refrigerators, and housing is related to the number of people in the country,
and the average wealth per person. The more people there are, the greater the
use. The wealthier the people, also the greater the use. Cultural and other
local geographic and climatic factors are also important, but population and
wealth explain much of the observed differences between countries.
This general idea of a product's use being related to population and
wealth applies not only to easily observable products such as cars and
refrigerators, but also to the things that are used to make these products.
For example, steel is used to make a large number of products, including cars
and refrigerators. The amount of steel used in different countries is
therefore related to the population and wealth of the country.
The use of CFCs, like steel, can also be thought of as being related to
the population and wealth of a country. Using a statistical technique call
"regression analysis," the manner in which the use of CFCs has varied in the
past with population and wealth can be estimated in the form of an equation.
These estimated equations can then be used to identify how CFC use may change
in the future as population and wealth change.
The equations estimated in this analysis relate historical CFC use per
person on the one hand to GNP per person and the price of CFCs on the other.
These equations are used with scenarios of future rates of growth in GNP per
person, the number of people, and the real price of CFCs to project the future
use of CFCs.
The advantages of this method are that it is simple and straightforward,
it can be implemented with the type of data that exist, and it can use
standard projection variables (population and GNP) to generate scenarios. The
primary shortcoming of this approach is that it does not describe the
underpinnings of the demand for CFCs. Instead, it implicitly assumes that the
structure of the demand for CFCs remains constant through time. If the
factors affecting demand change substantially, then the relationships
estimated using this approach will no longer be appropriate for projecting
future scenarios of CFC use. Examples of important changes in the factors
affecting demand include: changes in the relative prices of CFCs and their
-------
-6-
substitutes due to changes in the prices of the raw materials used to make the
products or changes in the technologies used to make the products; and changes
in the ability to use or handle substitutes due to findings regarding their
toxicity. Although it would be preferred to have available a rigorous demand
analysis that can reflect these potential changes, it was not undertaken
because:
• the data needed to perform the analysis was not
readily available; and
• for the variables needed to project demand,
established scenarios or estimates far into the future
are not unavailable, and would have had to be developed.
The lack of good estimates of the future values of the detailed determinants
of CFG demand is particularly important because projections of CFC use will
only be as good as the projections of the variables used to derive CFC use.
Using historical data on GNP, population, and prices (see Annex A for the
sources of the data) along with the historical data on CFC use, regression
analysis was used to estimate relationships for the following 10 combinations
of regions of the world; CFC type; and aerosol/non-aerosol application:
1. U.S.; CFC-11; aerosol
2. U.S.; CFC-11; non-aerosol
3. U.S.; CFC-12; aerosol
4. U.S.; CFC-12; non-aerosol
5. U.S.; CFC-22; total
6. U.S.; CFC-113; total
7. OECD-U.S.; CFC-11; aerosol
8. OECD-U.S.; CFC-11; non-aerosol
9. OECD-U.S.; CFC-12; aerosol
10. OECD-U.S.; CFC-12; non-aerosol.
Of note is that the analysis of U.S. aerosol applications only used data
through 1974. The post-1974 use was assumed to be influenced by the rising
concern over ozone depletion and the subsequent U.S. ban on the use of CFCs as
aerosol propellants. Similarly, the OECD-U.S. aerosol analysis only used data
through 1977. By truncating the time series in this manner, the resulting
relationships reflect the level of CFC aerosol use in these locations in the
absence of existing restrictions. Consequently, these relationships can be
used to assess whether the current restrictions are binding (i.e., the extent
to which use would be expected to have exceed the current levels absent the
attention given to the ozone-depletion issue and the subsequent regulations)
and to project scenarios of use in the absence of existing restrictions.
Results of the regression analysis are displayed in Annex B.
Due to a lack of data, relationships could not be developed for the
following region of the world, CFC combinations:
Non-OECD use of all CFCs (CFC-11; CFC-12; CFC-113;
CFC-22);
-------
• OECD-U.S. use of CFC-22 and CFC-113.
The next section describes how the relationships estimated in the regression
analysis are used to generate the scenarios. Also described are the methods
used to project potential use for the portions of the world for which
historical relationships could not be estimated.
2.3 Implementation of the Scenarios
Five scenarios of potential future use of CFCs were generated. Three
scenarios, Low, Medium, and High, report the likely range of future CFC use.
In addition, two bounding scenarios were generated: (1) "Limits to Growth," a
very low scenario; and (2) "No Limits to Growth," a very high scenario.
Although it is possible, it is unlikely that future use will fall outside
these bounding scenarios in the absence of government regulatory
intervention. For total CFC use in the U.S., and for CFC-11 and CFC-12 use in
the OECD, the scenarios were generated by inserting projections of population
and GNP per capita into the equations developed using the regression
analysis. However, as described in the previous section, equations could not
be developed for CFC-22 and CFC-113 use in the the non-LI.S. OECD countries and
for all the CFC use in the countries outside the OECD. Therefore, alternative
approaches were used to project scenarios for these areas.
All the approaches use population and GNP per capita as the basis for the
projections. Consequently, these data are described first. Then the manner
in which the regression equations were used is described. Finally, the
alternative approaches are presented.
Population and GNP Per Capita Projections
The potential future use of CFCs will be driven by the growth in the world
population and the expected wealth of that population. A variety of
projections have been published that provide estimates of potential future
population and GNP per capita through 2075. These estimates were reviewed
along with information describing historical rates of growth. Annex E
describes the data reviewed, and the assumptions used to create five
projections of population and GNP per capita through 2075.
Exhibit 2 summarizes the population projections for the world. The
highest projection, 13.6 billion people by 2075, is nearly double the low
projection of 7.1 billion by 2075. The growth rates implied in the
projections are also shown in the exhibit. Of note is that all the
projections have the general trend of slowing growth rates over time.
Relative to the historical rate of population growth from 1925 to 1975 (1.9
percent), all the projections start out with slower growth.
Not displayed in Exhibit 2 are regional differences in the population
projections. The developing countries are expected to grow more rapidly than
the developed countries in all the projections. Zero population growth (ZPG)
is assumed to be achieved in all regions in the Limits to Growth scenario by
-------
-8-
EXHIBIT 2
SUMMARY OF POPULATION PROJECTIONS
(Billions of people -- rates of growth in percent per year)
SCENARIO
1985
2000
2025
2050
2075
Limits to
Growth
4
(1
5
(0
6
(0
7
(-0
.6
.1%)
.4
.8%)
.5
.5%)
.3
. 1%)
Low
4.7
(1.5%)
5.9
(0.9%)
7.4
(0.2%)
7.7
(0.1%)
Medium
4.7
(1.5%)
5.9
(0.9%)
7.4
(0.4%)
8.2
(0.1%)
High
4.8
(1.6%)
6.1
(1.1%)
8.2
(0.6%)
9.5
(0.2%)
No Limits
to Growth
5
(1
6
(1
9
(1
12
(0
.0
.8%)
.5
.5%)
.5
.0%)
.1
.5%)
7.1
7.9
8.5
10.0
13.6
Source: See Annex E.
2050. The U.S. and other OECD countries achieve ZPG by 2075 in the Low and
Medium scenarios. Annex E presents the regional breakdowns of the projections,
Exhibit 3 displays the diversity among the GNP per capita projections. By
2075, the highest projection is over five times larger than the lowest
projection. The range of uncertainty in the economic growth projections is
larger than the range for the population projections. Compared to the
historical rate of growth from 1925 to 1975 (2.1 percent per year), the
projections start out slower.
Scenarios Based on the Statistical Analysis
The statistical analysis produced 10 equations that were used to generate
scenarios for 10 region, compound, end use combinations (see Section 2.2 and
Annex B). The scenarios were developed from each of the 10 equations by: (1)
varying the population and GNP per capita projections; and (2) modifying the
regression coefficients to reflect uncertainty.
The population and GNP per capita projections discussed above were used.
For example, the Low CFG scenario is based on the low population and low GNP
per capita projections. The differences among the population and GNP per
capita projections account for the majority of the differences in the CFC use
projections across the scenarios (over 80 percent).
-------
-9-
EXHIBIT 3
SUMMARY OF GNP PER CAPITA PROJECTIONS
(1975 U.S. $ -- rates of growth in percent per year)
SCENARIO
1985
2000
2025
2050
2075
Limits to
Growth
1,
^
?
*™ »
(0
2,
(0
2,
(0
900
.0°,)
200
.5?;)
500
.5%)
850
.3%)
Low
1,900
(1.5%)
2,400
(1.2%)
3,200
(1.0%)
4,100
(1.0%)
Medium
1
(
2
>
1
»
(1
3
(
5
(
t
1
)
1
900
.7%)
450
.7%)
750
.7%)
700
.7%)
High
1
(1
2
(2
4
(2
6
(2
,950
.9%)
,550
.0%)
,200
.0%)
,883
.0%)
No
to
1
(2
2
(2
5
(2
9
(2
Limits
Growth
,950
.3%)
,750
~o/ \
,100
.5%)
,450
.5%)
3,050
5,250 8,650
11,242
17,550
Source: See Annex E.
The uncertainty in the estimates of the regression coefficients was also
reflected in the scenarios. Regression analysis provides estimates of the
expected coefficients in the equations. There is uncertainty in these
estimates, which is expressed as the standard error of the estimated
regression coefficient. The larger the standard error relative to the size of
the coefficient, the more uncertain the estimate of the coefficient (and hence
the relationship).2 To reflect this uncertainty, the coefficient on GNP per
capita was varied across each of the scenarios. Changing the coefficient by
plus and minus one standard error of the estimate changed the scenarios by
approximately 10 to 20 percent in each direction. These changes are small
relative to the differences caused by the alternative population and GNP per
capita projection. The Low, Medium, and High scenarios presented below are
based on the estimate of the coefficient that is produced by the regression
analysis. The Limits to Growth scenario is based in this estimate
2 It should be noted that the estimates of the standard errors are
probably biased downward as the result of serial correlation of the data.
Consequently, the relationships are more uncertain than indicated by the
regression results. For more information on the influence of serial
correlation on uncertainty estimates in regression analysis see: Pindyck,
R.S. and D.L. Rubinfeld, Econometric Models and Economic Forecasts,
McGraw/Hill (New York), 1976, pp. 106-108.
-------
-10-
minus one standard error of the estimate. The No Limits to Growth scenario
is based on this estimate plus one standard error of the estimate.
Several additional adjustments were performed for all the scenarios. To
reflect the current ban on the use of CFCs as an aerosol propellant in
non-essential applications in the U.S., the annual aerosol use of CFC-11 and
CFC-12 was not permitted to exceed 4.5 million kilograms each. Because the
aerosol use in the U.S. would exceed this level without the ban, this
adjustment sets U.S. aerosol use throughout the period examined at this
level. Similarly, the aerosol use restrictions in other OECD countries were
reflected. Because the European Economic Community (EEC) countries have
agreed to reduce their aerosol use to 70 percent of their 1976 use, the EEC
use had to be projected separately. To do so, the non-U.S. OECD estimates
were divided into an EEC portion and a non-EEC portion using a constant
proportion. The EEC portion is set at approximately 71 percent of the aerosol
uses of CFC-11 and CFC-12 in the non-U.S. OECD (based on data from the
1970s). To reflect the EEC aerosol restrictions, the annual EEC use of these
CFCs in aerosol applications is limitted to 54.5 million kilograms and 69.4
million kilograms, respectively.
Other non-EEC countries have also instituted controls on the use of CFCs
in aerosol applications. Exhibit 4 shows the countries and the percent
reductions achieved to date. As shown in the exhibit, the controls reduced
total use by approximately 40 percent. This reduction is reflected in the
scenarios by assuming that future use would only be 60 percent of what it
would be estimated to be without the controls.
EXHIBIT 4
REDUCTIONS IN CFC-I! AND CFC-12 USE IN AEROSOL
APPLICATION IN NON-U.S., NON-EEC OECD COUNTRIES
Country
Australia
Austria
Canada
Finland
Japan
New Zealand
Norway
Sweden
Switzerland
Percent CFC Use as Percent of
Reduction Non-U.S., Non-EEC OECD
35 16
30 5
79 21
20 5
25 21
49 5
100 2
100 5
23 11
Weighted
Percent Reduction1
6
1
16
1
5
2
2
5
_2
TOTAL: 40 percent
1 Derived by multiplying first two columns.
Source: Derived from: Report on Chlorofluorocarbons, OECD, April 1982.
-------
-11-
Scenarios Based on Alternative Approaches
Because equations based on historical data could not be devloped for some
CFC uses in portions of the world, alternative approaches were used to
generate scenarios for these regions. The scenarios for CFC-22 and CFC-113
use in the non-U.S. OECD countries were developed by assuming that the use of
these CFCs in the U.S. would remain a constant fraction of the total use in
the OECD. For example, recent data indicate that the U.S. share of total
CFC-113 use in the OECD is approximately 65 percent.3 From the regression
equations, scenarios of the potential future U.S. use of CFC-113 are
generated. To generate the scenarios for CFC-113 in the other OECD countries,
it is assumed that the 65 percent U.S. share remains constant, meaning that
the use of CFC-113 in the non-U.S. OECD countries is approximately 54 percent
of the estimated U.S. use." The U.S. share of OECD use of CFC-22 is
approximately 80 percent.
The implications of this approach is that the scenarios for non-U.S.
CFC-22 and CFC-113 use are based on projections for the U.S. To the extent
that the U.S. data reflect the likely growth rate of the total use of these
CFCs in the OECD, the method provides a reasonable estimate of future OECD
use. However, the share of OECD use accounted for by the U.S. may change in
the future, probably becoming smaller as the other OECD countries capture
larger shares of the electronics markets. Consequently, even if the total
OECD CFC scenarios are reasonable, the U.S. share of this use may be biased
upward.
Scenarios could not be based on historical data for the non-OECD
countries. Instead, for purposes of projecting this use, it was assumed that
per-capita use would increase over time at a rate described by an "S-shaped"
curve. This type of curve, shown in Exhibit 5, reflects the hypothesis that
as wealth per capita grows over time, the use of CFCs will also grow, but not
in a linear fashion. Instead, once market penetration reaches a threshold (in
the exhibit shown as time T to T9) the growth in CFC use may exceed the
growth in wealth. Thereafter, the rate of increase in the use of CFCs
decreases relative to increases in wealth, as CFCs are used more efficiently
or as markets become saturated.
3 Based on a comparison of U.S. data with OECD data. U.S. data from:
A.D. Little, Preliminary Economic Impact Assessment of Possible Regulatory
Actions to Control Emissions of Selected Halocarbons. September 1976. OECD
data from: OECD (1981), Summary of the Conclusions Regarding Scenarios of
the Ad Hoc Meeting. ENV/Chem/PJC/81.79.
* The non-U.S. OECD share of the OECD use is 35 percent. This quantity
is 35 * 65 = 54 percent of the projected U.S. use.
-------
-12-
EXHIBIT 5
EXAMPLE OF AN S-SHAPED CURVE
Final Use
Use/person
Current Use/person
TIME
To specify this S-shaped curve, the logistic function was used.5 The
shape of the curve is fully defined by specifying the current use per capita
at time T and the final use per capita at time T^. Current use per
capita estimates are shown in Exhibit 6. For example, in the non-U.S. OECD,
current use per capita of CFC-11 is approximately 0.58 kilograms. In the
non-OECD countries, use per capita is currently approximately 1.2 percent of
this value (or about 0.007 kilograms). Using the logistic function, this
fraction of 1.2 percent will grow over time to the final fraction specified
for 2075.
5 The logistic function is f = l+exp(-a-bt); where
are parameters defining the shape of the curve.
"t"
is time and 'a
it/-u
and "b" are parameters defining the shape of the curve. The value for 'f
computed for each time, t, is multiplied by ]th& expected finapuse to compute
the use at time t. For example, the final (i.e., ending year) use is defined
as the use per capita expected to be achieved in 2075. To estimate the use
per capita in 2050, the value of "f" in 2050 (which will be less than 1.0) is
multiplied by the final per-capita/ use.
-------
-13-
EXHIBiT 6
CURRENT PER-CAPITA USE OF CFCs
CFC
Annual
Per-Capita Use
in OECD Countries
Other than the U.S.1
(kg/person)
Annual Per-Capitalf Use
in Non-OECD Countries
as a Percentage
of Annual Per-Capita
Use in OECD Countries
CFC-11
CFC-12
CFC-22
CFC-113
0
0
0
0
.58
.58
.06
.04
1
5
3
3
.22
.32
.83
.83
1 Estimated by taking the most current estimate of the total use
available (1982 or 1983) and dividing by the total number of people in the
OECD.
2 Current total use for CFC-11 and CFC-12 was taken from "Production,
Sales, and Calculated Release of CFC-11 and CFC-12 through 1982," Chemical
Manufacturers Association (1983). The CMA cautions that these estimates are
uncertain because they are based on extrapolation of data reported for earlier
years. However, estimates of the current use in the non-OECD countries do not
influence the overall scenarios significantly.
3 Current total use of CFC-22 and CFC-113 was derived using estimates of
the share of total world use calculated for these countries from data in OECD
(1976), Fluorocarbons: An Assessment of Worldwide Production, Use and
Environmental Issues, pp. 30-35.
-------
-14-
Th e final fraction was estimated using the projected GNP per capita in the
non-OECD countries, and the current relationship between per capita use and
per capita GNP in the non-U.S. OECD countries. For example, in the medium
population and GNP per capita projection, the GNP per capita in the non-OECD
countries reaches approximately 85 percent of the 1980 GNP per capita in the
non-U.S. OECD countries. Consequently, it was assumed that by 2075, CFC use
per capita in the non-OECD countries would reach this same fraction (85
percent) of the 1980 use per capita in the non-U.S. OECD. The estimate for
the low and high projections were 53 percent and 111 percent.
Exhibit 7 summarizes the assumptions in the scenarios.
3. RESULTS
This section presents five scenarios of future CFC use and emissions
worldwide. These scenarios incorporate the current restrictions on CFC use in
aerosol applications in the U.S. and elsewhere.
Exhibit 8 summarizes the results for the scenarios.6 For example, the
Medium scenario for CFC-11 shows world use growing from approximately 400
million kilograms in 1985 to about 7,600 million kilograms by 2075. This is
an average rate of growth of 3.3 percent per year. The growth rate in the
Medium scenario is not constant, however. The early years have nearly a 4.5
percent annual rate of growth, and the later years have growth at 2.0 percent
per year.
The annual use of CFC-11 in the Low scenario in Exhibit 8 grows on average
at only 2.6 percent per year throughout the entire period examined, resulting
in 4.3 billion kilograms of use by 2075. The High scenario has larger use of
CFC-11 over time, with an average annual growth rate of 3.7 percent.
The growth in use per capita is smaller than the growth in total use
because population also increases over time. For the total of all four CFCs
examined, world per capita use reaches the following levels in the five
scenarios (smallest to largest): (1) 0.65 kilograms; (2) 1.3 kilograms; (3)
2.1 kilograms; (4) 2.8 kilograms; and (5) 4.6 kilograms. Current per capita
use in the non-U.S. OECD countries is approximately 1.3 kilograms, and in the
U.S. is approximately 1.6 kilograms (both these estimates reflect existing
restrictions on aerosol applications). Consequently, per capita use in the
1.3 to 2.8 kilograms range from the Low, Medium, and High scenarios seems
plausible. The Limits to Growth scenario has per capita use well below
current rates in the developed world. Non-OECD countries would be required to
have very low economic and population growth over the next 90 years to achieve
the per capita use estimates in the Limits to Growth scenario. Similarly, the
No Limits to Growth Scenario has very high per capita as rates, requiring both
large population growth and excellent economic performance to be achieved.
Annex D presents scenario results for each of the major regions of the
world.
-------
EXHIBIT 7
SUMMARY OF SCENARIO ASSUMPTIONS
PopnI at ion
Project ion
CNP Per
Cap ila
Project ion
ciowth of
Use in OECD
Countries
Limits to
Growth Scenario
(Grows at an average
j rate or 0.5 percent
I ( 1985-2075)
Grows at an average
rate of 0.58 percent
(1985-2075)
IGrows at average real
irate of 0.5 percent
|(1985-2075)
Grows at average real
'rate of I.I percent
(1985-2075)
(Historical re lationship Based
|reduced to reflect un-
Icertainty (minus one
_ [standard errorj
"I
Growth of Use I Follows ;m S-shaped
in Non-OECO (curve. Reaches 30
Countries (percent of current per
(capita use in non-U.S,
(OECD countries
Restr ict ions
Performed with current
aerosol use restric-
t ions
Low Scenario
on h i stor icaI
relat ionship
Follows an S-shapod
curve. Reaches 53
percent of current per
cap i ta use i n i
OECD countries
Performed with
aerosol use res
t ions
je
ant
rea 1
nt
a 1
;d
3
t per
-U.S.
rrent
ic-
Medium Scenario
Grows at an average
rate of 0.65 per-
cent ( 1985-2075)
Grows at an average
rea 1 rate of 1.7
percent ( 1985-2075)
Based on h i stor ica 1
re lat ionsh i p
Follows an S-shaped
curve. Reaches 85
percent of current
per cap i ta use i n
rion-U.S. OECD
countries
High Scenario
Grows at an average
rate of 0.8 percent
( 1985-2075)
Grows at average real
rate of 2.0 percent
( 1985-2075)
Based on h i stor ica 1
re lat ionsh i p
N
Gr
Grows
rate
( 1985
Grows
ra te
( 1985
H i sto
i nc re
uncer
stand
Fo 1 lows an S-shaped Fo 1 lo
curve. Reaches III jcuive
percent of current perjperce
capita use in rion-U. S. j cap i t
OECD countries (OECD
Performed with cur- Performed witfi cur-
rent aerosol use rent aerosol use
restrictions restrictions
Perl o
aeros
t ions
No limits to
Growth Scenario
Grows at average real
rate of 2.5 percent
a I re I a tionship
J to retlect
i ty (plus one
erro r )
iws an S-shaped
Roaches IY?
nt of current per
a use in non-U.S.
count r i es
3d wi tri current
use rustric-
-------
EXHIBIT 8
SUMMARY OF CFC SCENARIOS
LOW SCENARIO
MEDIUM SCENARIO
HIGH SCENARIO
CFC- 1 1
CFC- 12
CFC-22
CFC-I 13
CFC-I 1
CFC-12
CFC-22
CFC- 1 13
CFC-I 1
CFC-12
CFC-22
CFC- 1 1 3
Potential Future Use
(millions of kilograms)
1985
2000
2025
2050
2075
Average Growth
1985-2000
2000-2025
2025-2050
2050-2075
«400
725
1,1475
2,550
14,325
Rate (%)
4.0
2.9
2.2
2.2
550
950
1,775
2,750
14, 175
3.7
2.6
1 .7
1.7
1 10
220
380
5*40
830
14.7
2.2
1 .>4
1 .7
95
200
3')0
1475
725
5.0
2.2
1 .14
1 . 7
400
785
2,035
14,625
7,650
14.5
3.9
3.3
2.0
560
1,065
2,525
14,825
7,350
14.14
3.5
2.6
1 .7
1 10
235
500
930
1,550
5. 1
3. 1
2.5
2. 1
95
210
«450
820
1 ,380
5.4
3. 1
2.5
2. 1
»430
980
3,850
8,2140
1 1,680
5.6
5.6
3. 1
1 .14
600
1 , 1450
14,620
8, 130
1 1,2140
6.0
14.7
2.8
1 .3
1 15
280
790
1 ,1480
2,1425
6.2
14.2
2.6
?.o
95
250
6/5
1 ,280
2, 170
6. '4
14. 1
2.6
2. 1
Source: ICF Incorporated Estimate.
-------
[XHIIilT 8 (continued)
SUMMARY OF CFG SCENARIOS
LIMIIS 10 GROW III SCENARIO
NO LIMITS 10 GROW IH SCENARIO
Potential Future Use
(mill ions of ki log rams)
1985
2000
2025
2050
2075
Average Growth Rate (%)
1985-2000
2000-2025
2025-2050
2050-2075
CFC-I 1
350
550
860
1 ,'lOO
2,050
3.2
1.9
1.9
1.6
CFC- 12
*460
700
1 , OUO
1 .520
2,000
2.8
1 .6
1 .5
1 .0
CFC-22
75
135
180
250
320
U. 1
1 .2
1 . 3
1 .0
CFC- 1 13
75
!30
175
230
290
t». 1
1 . 1
1 . 1
1 .0
CFC- 1 1
500
1,325
6,550
16, 100
25,850
6.5
6.6
3.7
1 .9
CFC-12
800
2, 120
8,020
16,075
25,250
6.7
5.5
2.8
1 .8
CFC-22
160
1*20
1 ,*400
3, 100
5,885
6.7
5.0
3.2
2.6
CFC- 1 13
125
3»45
1 , 150
2,570
5,075
7.0
U.9
3. 3
2.8
Source: ICF Incorporated Estimate.
-------
-18-
Also of interest is the projected "intensity of use" in each of the
scenarios. Intensity of use measures the contribution that CFCs are expected
to make to overall economic activity, measured in kilograms per 1,000 dollars
of GNP (1975 U.S. dollars). In 2075 the intensity of use in each of the five
scenarios is estimated as: (1) 0.21 kilograms; (2) 0.24 kilograms; (3) 0.24
kilograms; (4) 0.25 kilograms; and (5) 0.26 kilograms. The current intensity
of use in the U.S. and in the non-U.S. OECD countries is approximately 0.2
kilograms. The scenarios, therefore, do not reflect an unreasonably large
intensity of use of CFCs over the next 90 years.
4. LIMITATIONS
This section discusses the major limitations of the data and methods used
to project scenarios of CFC use and emissions. Many of these limitations are
inherent in all projections far into the future. Others, however, relate to
the lack of historical data describing CFC use in many parts of the world and
the simple methods used to generate the scenarios. Individuals considering
using the scenarios should be fully aware of the limitations presented here.
4.1 Limitations in Data
Historical data on CFC use and emissions form the basis for the
projections of the CFC scenarios. The data are most detailed and complete for
the United States, from 1970 to the present. Aggregate data (not
differentiated by CFC or use) are available for the OECD for CFC-11, CFC-12,
and CFC-22. However, data are particularly sketchy in the following areas:
• Prior to 1970, the division of CFC-11 and CFC-12
between aerosol and nonaerosol uses in the United States
is not documented. The division was made by assumption
for purposes of this study.
• Data describing the historical use of CFC-113 outside
the United States have not been developed. Rough
estimates of the current U.S. share of OECD use of
CFC-113 were the only data available to describe the use
of CFC-113 outside the U.S.
• The mix of uses of CFC-11 and CFC-12 in the OECD is
only documented for selected years (i.e., the fraction
used in foams, air conditioners, aerosol propellants,
etc.). The mix of uses in the United States was assumed
to apply to the entire OECD for purposes of developing
emissions rates from use.
• Data for non-OECD countries are unreliable or
nonexistent.
These data difficulties were overcome by adopting reasonable assumptions.
However, alternative assumptions are plausible that may influence the
scenarios.
-------
-19-
4.2 Limitations in Methods
The methods used in this study to project scenarios of future CFG use are
simple and straightforward. Historical data are used to develop relationships
between CFC use and population and GN'P. These relationships are then used as
the basis for projecting future scenarios. The limitations of this method
include:
• The structural underpinnings of demand are not
estimated. The method assumes that the determinants of
demand, in relation to the variables used to make
projections (population and GN'P), remain unchanged.
• The relationships are based on a fairly narrow range
of experience, from a unique history. In the future,
population and GNP will be far outside the historical
range for the OECD countries. Therefore, even if the
relationships among the variables are estimated
correctly, there is considerable uncertainty surrounding
estimates of future use.
• Estimates of population and GNP far into the future
are inherently uncertain. The range of estimates used
may understate the full range of possible future
outcomes.
• Non-OECD use of each CFC was projected by assuming
that per-capita use in these countries would reach a
portion of current per-capita use in the non-U.S. OECD.
Of note is that the OECD experience may not be
applicable to these non-OECD countries due to
differences in culture, economies, climate, and other
factors.
• By using historical data, the scenarios reflect the
historical rate of the introduction of new uses for
CFCs. However, the method does not explicitly explore
the potential for new uses, nor does it address the
potential for new substitutes or more efficient use of
CFCs in existing uses.
• Constraints in the future production of CFCs were not
explored.
These limitations can be addressed by performing additional research and
analysis. However, scenarios that project far into the future will always be
fairly uncertain. The next section provides recommendations for next steps.
-------
-20-
4,3 Next Steps
To improve the scenarios reported here, analysis should focus on potential
natural or man-made constraints to future CFC production and use. This
analysis could first explore the implications of the scenarios of future CFC
use in termes of demand for complementary products (e.g., automobiles,
housing, furniture). Other issues include:
• the likely future availability of the raw materials
needed to make CFCs;
• the capital investment required in plant and equipment
needed to produce CFCs at the projected rate;
• the potential elimination of CFC substitutes in
certain applications due to their toxicity (e.g., carbon
tetrachloride used as a solvent in developing nations);
• increased refrigeration and space-cooling requirements
as global warming occurs; and
• differences in potential future use among countries
developing at different rates (e.g., rapidly developing
countries like Taiwan, Brazil, Korea), and among
countries that may follow very different development
paths (e.g., Tanzania).
By examining these factors, the potential constraints on future CFC use can be
explored.7 These constraints may indicate when important turning points in
the scenarios occur, i.e., when use cannot grow at rates indicated by the
analysis of the historical data.
If constraints are found to be important, the structural underpinning of
demand should also be explored. This information will be required to assess
how the future constrained amounts of CFCs will be divided among potential
competing uses. Additionally, this information may be useful for improving
estimates of future uses in the non-OECD countries by identifying important
differences between these countries and the OECD countries.
7 Of note is that as recently as 1976 it was reported that the projected
supply of fluorine in the world is "immense and provides no practical limit to
CFM [CFC] production," in Halocarbons: Effects on Stratospheric Ozone,
National Academy of Sciences, 1976, p. 174. Consequently, the availability
and/or price of this raw material may not pose a constraint on the future
production of CFCs. More recent analyses confirm this conclusion, see: Mooz,
William E., Kathleen A. Wolf, Frank Camm, "Potential Constraints on Cumulative
Global Production of Chlorofluorocarbons," The RAND Corporation, WD-2864-EPA,
December 1985, and Gibbs, Michael J. and Robin Weiner, "Assessment of the
Availability of Fluorine for the Production of Chlorofluorocarbons," ICF
Incorporated, February 1986.
-------
A-l
ANNEX A
This Annex presents the historical data used to project scenarios of
future CFC use. The data are presented in a series of tables. Each table is
followed by a description and the footnotes that list the sources of the
data. Exhibit A-l lists the tables presented.
-------
A-2
EXHIBIT A-1
LIST OF TABLES IN ANNEX A
Table Title Page
A-1
A-2
A-3
A-4
A-5
A-6
A-7
A-8
U.S. CFC Production Data - Total
U.S. CFC Production Data - Aerosols
U.S. CFC Production Data - Non-Aerosols
Nominal and Real CFC Prices
U.S. Economic Data
OECD CFC Production Data
OECD Economic Data
Projections of GNP and Population Relative to 1975
A-3
A-5
A-7
A-9
A-ll
A-13
A-17
A-19
-------
A-3
TABLE A-1
U.S. CFC PRODUCTION DATA - TOTAL
(Millions of Kilograms)
YEAR
CFC-11 a/
CFC-12 a/
CFC-22 a/
CFC-113 c/
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
22.90
27.40
32.80
41.20
56.60
63.60
67.40
77.30
77.30
82.70
92.70
108.20
110.90
117.00
135.90
151.40
154.70
122.30
116.20
96.40
87.90
75.80
71.70
73.80
63.70
73.10
59.60
71.30
75.50
78.70
94.30
98.60
103.40
123.10
129.90
140.50
147.70
166.80
170.30
176.70
199.20
221.70
221.10
178.30
178.30
162.50
148.40
133.30
133.80
147.60
117.00
134.30
15.10 b/
16.60
18.20
20.50
22.30
24.50
26.80
29.10
31.80
35.50
39.10
42.70
45.50
50.90
55.90
61.80
64.10
59.80
77.00
81.40
93.30
95.60
103.20
114.20
79.00
106.90
0.00
0.00
2.00
3.00
4.00
4.50
5.40
6.40
7.30
9.50
11.40
13.60
16.40
19.50
22.70
26.80
29.00
31.00
37.10
45.20
51.20
58.80
45.90
48.20
50.00
52.70
a/ See attached footnotes.
-------
A-4
TABLE A-1: U.S. CFC PRODUCTION DATA - TOTAL
Description:
Footnotes:
Table A-1 presents time series data on the total U.S.
production of CFC-11, CFC-12, CFC-22, and CFC-113 for the years
1958 to 1983. The primary source of these data is the U.S.
International Trade Commission. It was necessary to estimate
some values using assumptions based on actual data. The
footnotes which accompany the table describe these assumptions.
a) U.S. ITC. Synthetic Organic Chemicals. Annual Series.
These include CFC-22 used as an intermediate.
Approximately 28 percent of CFC-22 is used in the
production of teflon and other products.
b) Estimates for 1958-59 for CFC-22 are based on average
annual percent changes over the period 1960-65.
c) For 1958 to 1962, U.S. production is assumed to be 100
percent of total world production reported in: OECD, 1981,
Summary of the Conclusions Regarding Scenarios of the Ad
Hoc Meeting, ENV/CHEM/PJC/81.79, p. D.I. U.S. production
in 1963 was assumed to be 90 percent of the world total
reported in the OECD document. Data for 1964 to 1974 are
from: A.D. Little, Preliminary Economic Impact Assessment
of Possible Regulatory Actions to Control Emissions of
Selected Halocarbons, September 1976. Data for 1975 to
1979 are estimated by taking the portion of world CFC-113
production accounted for by the U.S. in 1974 (64.6 percent)
and multiplying by the world production reported in the
OECD document. The years 1980 to 1983 are projecrions
reported in U.S. EPA, Economic Implications of
Regulating Chlorofluorocarbon Emissions from Nonaerosol
Applications, (EPA-560/12-80-001), October 1980, pp. 67-68.
-------
A-5
TABLE A-2
U.S. CFC PRODUCTION DATA - AEROSOLS
(Millions of Kilograms)
YEAR CFC-11 CFC-12
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
19.90 a/
23.10
26.70
32.40
43.20
46.60
47.50
52.40
50.20
51.50
55.20
61.40
60.00 c/
63.10
72.10
83.40
86.70
73.30
64.00
36.50
32.90 d/
9.50
9.50
9.50
35.80 b/
42.80
45.30
47.20
56.60
59. 10
62.00
73.90
77.90
84.30
88.60
100.00
102.10 c/
106.20
117.50
135.20
123.90
94.50
80.30
47.30
41.90 d/
5.30
5.80
5.60
a/ See attached footnotes.
-------
A-6
TABLE A-2: U.S. CFC PRODUCTION DATA - AEROSOLS
Description:
Footnotes:
Table A-2 presents time series data on the total U.S. CFC
production of CFC-11 and CFC-12 for aerosol applications. The
statistical analyses performed to project potential aerosol
applications in the absence of government regulation used data
for the years 1958 to 1974. Use in the years following 1974
may be influenced by expectations regarding government
regulation.
a) Estimates for 1958-1969: based on the assumption that in
1958 the U.S. dominates the OECD market as CFC-11. In
1958, the CMA data (CMA. Production, Sales, and Calculated
Release of CFC-11 and 12 through 1982; Expanded Data) show
a market share of open cell, aerosol, and all other
production to be 87 percent of total CFC-11 production.
Thus, the U.S. aerosol share of total CFC production is
assumed to be 87 percent in 1958, declining smoothly to 54
percent in 1970, where 54 percent is the aerosol proportion
in 1970 described in footnote C. Total CFC-11 production
is based on U.S. ITC data.
b) Estimates for 1958-1969: based on the assumption that
CFC-12 aerosol production was 60 percent of total CFC-12
production, as it was in 1970.
c) Data for 1970-1977: U.S. EPA. Regulating
Chlorofluorocarbon Emissions: Effects on Chemical
Production. fEPA-560/12-80-0016) October 1980, pp. 73, 74,
80. Reports production figures from pp. 73-74, using
aerosol/nonaerosol proportions from p. 80.
d) Estimates for 1978 to 1981 computed by subtracting usage in
non-aerosol applications from total CFC usage. See Tables
A-l and A-3.
-------
A-7
TABLE A-3
U.S. CFC PRODUCTION DATA - NON-AEROSOLS
(Millions of Kilograms)
YEAR
CFC-11
CFC-12
CFC-22 f/
CFC-113 h/
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
3.00 a/
4.30
6.10
8.80
13.40
17.00
19.90
24.90
27.10
31.20
37.50
46.80
50.80 c/
54.00
64.00
68.00
68.00
49.00
52.20
59.90
55.00 d/
66.30
62.20
64.30
63.70 e/
73.10
23.80 b/
28.50
30.20
31.50
37.70
39.50
41.40
49.20
52.00
56.20
59.10
66.80
68.00 c/
70.80
81.60
86.60
97.10
83.90
98.00
115.20
106.50 d/
128.00
128.00
142.00
117.00 e/
134.30
15.10 g/
16.60
18.20
20.50
22.30
24.50
26.80
29.10
31.80
35.50
39.10
42.70
45.50
50.90
55.90
61.80
64. 10
59.80
77.00
81.40
93.30
95.60
103.20
114.20
79.00
106.90
0.00
0.00
2.00
3.00
4.00
4.50
5.40
6.40
7.30
9.50
11.40
13.60
16.40
19.50
22. 70
26.80
29.00
31.00
37.10
45.20
51.00
58.80
45.90
48.20
50.00
52.70
a/ See attached footnotes.
-------
A-8
TABLE A-3: U.S. CFC PRODUCTION DATA - NON-AEROSOLS
Description:
Footnotes:
Table A-3 presents time series data on U.S. CFC production for
non-aerosol applications. The table contains actual data for
CFC-11 and CFC-12 for the years 1970-1977; CFC-22 for the years
1960-1983; and CFC-113 for the years 1965-1974. The remaining
estimates are based on assumptions detailed in the footnotes.
a) Estimates for 1958-1969: CFC-11 non-aerosols are defined
as CFC-11 total minus CFC-11 aerosol (see Table A-2).
b) Estimates for 1958-1969: based on the assumption that
CFC-12 nonaerosol production was 40 percent of total CFC-12
production, as it was in 1970.
c) Data for 1970-1977: U.S. EPA. Regulating
Chlorofluorocarbon Emissions: Effects on Chemical
Production. (EPA-560/12-80-0016) October 1980, pp. 73,
74, 80. Reports production figures from pp. 73-74, using
aerosol/nonaerosol proportions from p. 80.
d) Estimates for 1978-1981: OECD. Economic Aspects of CFC
Emission Scenarios prepared by the Chemical Groups and
Management Committee, 15 September 1982, p. 32. Based on
the assumption that the split between CFC-11 and CFC-12 in
these years was the same as it was in 1977.
e) Estimates for 1981-1982: based on the assumption that
production for aerosol applications is zero, which makes
nonaerosol production equal to total production.
f) U.S. ITC. Synthetic Organic Chemicals. Annual Series.
g) Estimates for 1958-1959: based on the average annual
percent change for CFC production for the period 1960-1965.
h) For 1958 to 1962, U.S. production is assumed to be 100
percent of total world production reported in: OECD, 1981,
Summary of the Conclusions Regarding Scenarios of the Ad
Hoc Meeting, ENV/CHEM/PJC/81.79, p. D.I. U.S. production
in 1963 was assumed to be 90 percent of the world total
reported in the OECD document. Data for 1964 to 1974 are
from: A.D. Little, Preliminary Economic Impact Assessment
of Possible Regulatory Actions to Control Emissions of
Selected Halocarbons, September 1976. Data for 1975 to
1979 are estimated by taking the portion of world CFC-113
production accounted for by the U.S. in 1974 (64.6 percent)
and multiplying by the world production reported in the
OECD document. The years 1980 to 1983 are projections
reported in U.S. EPA, Economic Implications of
Regulating Chlorofluorocarbon Emissions from Nonaerosol
Applications, (EPA-560/12-80-001), October 1980, pp. 67-68.
-------
(ABLE A-M
NOMINAI AND KfAL CIC PRICES
(cents/kg)
YEAR
1958
1959
1960
1961
1962
1963
1961
1965
1966
1967
1968
1969
1970
1971
19/2
1973
1971
1975
1976
1977
1978
1979
1980
1981
1982
1983
crc-i
Norn i na 1
16
18
18
16
11
11
11
11
11
11
39
11
11
39
11
5?
71
71
71
86
92
101
.20 a/
NA
.80
.80
.20
.00
.00
.00
.06
.80
.80
.60
.80
.80
.60
.80
.80
.80
.80
.80
NA
.00 d/
.00
.00
NA
NA
1
Real
69
71
70
65
61
60
59
57
52
50
15
15
13
39
39
15
59
56
53
52
51
53
.96 b/
NA
.03
.39
.13
.39
.16
.17
.10
.87
.61
.63
.71
.51
.60
.53
.88
.16
.52
.11
NA
.63
.56
.30
NA
NA
CIC-1
Noinj na 1
66.
68.
66.
66.
63.
63.
63.
61.
61.
59.
57.
55.
57.
55.
52.
52.
68.
90.
90.
88.
91.
105.
119.
oo a/
20
00
00
80
80
80
60
60
10
20
00
20
00
80
80
20
20
20
00
NA
00 d/
00
00
NA
NA
2 CTC-22
Re_al
199.
100.
96.
95.
90.
89.
87.
82.
80.
75
69!
63.
62.
57.
52.
19.
59.
71.
68.
62.
5?.
58.
60.
91 b/
89
07
20
36
02
67
81
25
13
30
37
55
29
80
93
26
71
16
83
NA
52
85
98
NA
NA
Nomina 1
156
151
151
151
117
115
136
131
138
136
132
123
1 12
112
107
99
123
173
156
119
135
1 10
192
.20 a/
.80
.00
.00
.10
.20
.10
.20
.60
.10
.00
.20
.20
.20
.80
.00
.20
.80
.20
.60
NA
.00 d/
.00
.00
NA
NA
Real
236.52 b/
221 . 56
221.16
222. 13
208.75
202.60
187.11
180.17
180.56
172.53
159.92
111.95
122.69
116.86
107.80
93.62
107.06
138. 17
118.03
106.82
NA
82.61
95.28
98.39
NA
NA
crc-m
lorn i na 1
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
88.0 c/
88.0
88.0
85.0
95.0
107.0
111.0
123.0
NA
13'4.0 d/
1 52 . 0
159.0
NA
NA
Rea 1
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
96.2 b/
91.7
88.0
80.1
82.5
85.0
86. 1
87.8
NA
82.0
85.2
81.5
NA
NA
a/ See attached footnotes.
NA = Not ava i(able.
I'l I R KIVIIW DRAM -- Do Not Quote or Citt: ***
-------
A-10
TABLE A-4: NOMINAL AND REAL CFC PRICES
Description:
Footnotes:
Table A-4 presents time series data on the nominal and real
prices of CFC-11, CFC-12, CFC-22 and CFC-113. The primary
source of this data is the SRI Chemical Economics Handbook.
a) Data for 1958-1977: SRI. Chemical Economics Handbook.
b) Real prices computed using implicit GN'P deflator.
c) Data for 1970-1977: U.S. EPA. Economic Implications of
Regulating Chlorofluorocarbon Emissions from Nonaerosol
Applications. (EPA-560/12-80-001) October 1980, p. 79.
d) 1979-1982: List prices from the Chemical Marketing
Reporter.
-------
A-ll
TABLE A-5
U.S. ECONOMIC DATA a/
Year
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
GNP
(bill. 1972 $)
680.9
721.7
737.2
756.6
800.3
832.5
876.4
929.3
984.8
1011.4
1058.1
1087.6
1085.6
1122.4
1185.9
1254.3
1246.3
1231.6
1298.2
1369.7
1438.6
1479.4
1475.0
1513.8
1485.4
1534.8
Population
(1000s)
174,882
177,830
180,671
183,691
186,538
189,242
191,889
194,303
196,560
198,712
200,706
202,677
205,052
207,661
209,896
211,909
213,854
215,973
218,035
220,239
222,585
225,055
227,704
229,849
232,057
234,249
GNP Per Capita
(1972 $/person)
3,890
4,060
4,080
4,120
4,290
4,400
4,570
4,780
5,010
5,090
5,270
5,370
5,295
5,405
5,650
5,920
5,830
5,700
5,940
6,220
6,465
6,575
6,480
6,587
6,400
6,550
Annual Growth Rate
of GNP Per Capita
f Of \
C/oJ
4.4
0.4
1.0
4.1
2.6
3.9
4.6
4.8
1.6
3.5
1.9
-1.4
2.1
4.5
4.8
-1.5
-2.2
4.2
3.8
3.9
3.0
-2.7
1.6
-2.8
2.3
a/ See attached footnotes.
-------
A-12
TABLE A-5: U.S. ECONOMIC DATA
Description: Table A-5 presents time series data on U.S. GNP (in billions
of 1972 dollars) and population (in thousands)
Footnotes: a) Economic Report of the President, 1984. pp. 222, 253
-------
TAEJI E A-6
OEC1) CFG PRODUCTION DATA
(Mill ions of Kg)
YEAR J CFC- 1 1
1958
1959
1960
1961
1962
1963
196U
1965
1966
1967
1968
1969
1970
1971
1972
1973
197'1
1975
1976
1977
1978
1979
1980
1981
1982
29.50
35.60
1(9.70
60.50
78. 10
93.30
111.10
122.80
1 <4 1 . 00
159.80
183. 10
217.30
238. 10
263.20
306.90
3'l9. 10
369.70
3 1 '4 . 1 0
339.80
320.50
308.90
289.50
289.60
286.90
282.60
Tola 1 Product ion
CFC-12
a/ 73.UO a/
87.60
99 . 140
108.50
128. 10
1 '46. 140
170. 10
190. Ill
216.20
2U2.80
267.50
297. 30
321. 10
3'41.60
379.90
1423.30
l4'42.80
381 .00
1410.70
382.80
372.10
357.20
350.20
351.30
328.00
CFC-22
7.60 b/
11.30
12.20
12.20
15.50
17.60
22.1)0
25.10
31.60
37.140
45. 70
56.00
58.80
6'4 . 60
70.10
76.60
87.90
73.70
80. 70 c/
88.140
96 . 80
106.00
116.10
127.10
139.20
A
CFC- 1 1
25. 70 d/
30.70
'42. 18
50 . 52
63.UO
73.63
85.90
92 . 50
102. 70
1 T4.53
127.50
1't7. 46
159.33
168.81
191 .81
211.142
222.53
183.79
195. 10 f/
16U. 70
143.60
111. 60
105. 10
9*4.00
92.50
Assumpt ion 1
it?. 140 d/
50.60
57.50
62.70
7 'i.OO
84.60
98.30
1 09 . 90
1 25 . 00
T40. 30
15U.60
171.80
185.60
197.140
219.60
2'4'4 . 70
255.90
220.20
Aerosol Producti on
t ion 1
0 d/
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Assumption 2
314.140 e/
145.50
53.70
59.00
74.20
87.20
1014.80
115.30
132.00
150.140
168.80
187.00
201.70
2114.30
236.60
263.00
276.10
229.80
Assumpt i on 3
I4U.OO g/
52.60
59.60
65.10
76.90
87.80
102. 10
1 1 '4 . 1 0
129.70
U45.70
160.50
1 78.140
192.70
205.00
2214. 10
258.20
2118.00
201 .90
* Estimates of the aerosoI/non-aerosoI division of CFC-12 for 1958 to 1975 were performed using three
sets of assumptions. See footnotes, d, e, and q for details.
-------
TAB!E A-6 (continued)
OECD CK; PRODUCT I ON DATA
(Mi I I ions of Kg)
YEAR
1958
1959
1 960
1961
1962
1963
1 964
1965
1966
1967
1968
I960
1970
19/1
972
973
974
9/5
976
977
978
979
1980
1981
1982
L _____
crc- i i
3.80 d/
4.90
7.60
9.98
14. 70
19.67
25.20
30.30
38.30
'45.27
55.60
69. 8'.
78. 11
9'l.39
1 15.09
137.68
1 '1 7 . 1 /
130.31
144. 70
155.90
165.40
1 7 7 . 90
18 ..60
193.00
190.20
As sump i ion I
31 .00 d/
37.00
41 .90
45.80
54.10
61.80
71 .80
80.20
91 .20
102.50
1 12.90
125.50
135.50
144.20
160.30
1 /8.60
186.90
160.80
Non-Aeroso 1
C. C-
As sumption 2
39.00 e/
42. 10
45. 70
49.50
53.90
59.20
65.30
74.80
84.20
92.40
98. 70
1 10. 30
119.40
127.30
143.30
160. 30
166. 10
151 .20
Product ion
2»
Assumption 3
29.40 g/
35.00
39.80
43.40
51 .20
58.40
68.00
76.00
86.50
97. 10
107.00
1 18.90
128.40
136.60
155.80
165. 10
194.80
152.40
_
-
-
-
-
-
-
-
-
-
1 /3.50 f/
196. 70
208.40
214.20
212.20
223.80
212. 10
CFC-22
7.60 b
1 1 . 30
12.20
12.20
15.50
17.60
22.40
25.10
31 .60
37.40
45. 70
56.00
58.80
64 . 60
70. 10
76.60
87.90
73. 70
80. 70 c
88.40
96.80
106.00
116.10
127. 10
139.20
a/ See attached footnotes.
* Estimates of the aerosoI/non-aerosoI division of CFC-I2 for I958 to I975 were
performed using three sets of assumptions. See footnotes, d, e, and g for details.
-------
A-15
TABLE A-6: OECD CFC PRODUCTION DATA
Description:
Footnotes:
Table A-6 presents time series data on OECD CFC production for
1958 to 1982. The principal source for these data are the
Chemical Manufacturers Association (CMA) estimates based on the
reports of 21 companies involved in CFC production throughout
the world. The assumption underlying the estimates is that
production by CMA reporting companies is used almost entirely
in the OECD countries. The procedures for dividing total
production of CFC-11 and CFC-12 into aerosol and nonaerosol use
are described in footnotes d, e and g.
a)
b)
CMA. Production, Sales, and Calculated Release of CFC-11
and CFC-1" Through 1982. 12 August 1983, Schedules 2, 3.
1958-197
Worldwid
OECD. Fluorocarbons: An Assessment of
c)
reduction, Use, and Environmental Issues.
1976, Scuedule 4. CFC-22 data are end use only and do not
include CFC-22 produced as an intermediate.
Estimates for 1976-1982: based on the assumption that the
percent change remains constant at 9.5 percent, which is
equal to the average of the annual percent changes over the
period 1970-1975.
d) CMA. 12 August 1983. Estimates for 1958-1975: based on
the 1976 proportions for aerosol/nonaerosol reported in
schedules 5 and 6:
• CFC-11 production for aerosol is based on CMA
estimates assuming that all of the open cell, aerosols,
and all other production category was used for aerosols
in 1958. It is also assumed that in 1976, aerosols
make up 76 percent of this category, based on 1976
proportions for aerosols found in schedules 5 and 6.
The aerosol share of total CFC-11 production in this
time period (1958-1976) is thus 87 percent in 1958,
declining smoothly to 57 percent of total CFC-11 in
1976. The non-aerosol share of CFC-11 is assumed to be
total production minus aerosol production.
• CFC-12 production for aerosol =57.8 percent of total
CFC-12 production for nonaerosol =42.2 percent.
e) CMA. 12 August 1983. Estimates for 1958-1975: based on
the 1976 proportions for aerosol/nonaerosol reported in
schedules 5 and 6:
• CFC-12 for aerosol = 82.8 percent of CFC-12 for
aerosol + open cell + all other
-------
A-16
f) CMA. 12 August 1983. Estimates for 1976-1982: based on
assumption that aerosol sales = aerosol production.
g) Estimates for 1958-1975: based on the assumption that OECD
division between aerosol and nonaerosol use will follow the
U.S. experience (until the time of the U.S. ban on aerosol
production).
-------
A-17
TABLE A-7
OECD ECONOMIC DATA a/
Annual Growth Rate
GDP
YEAR (Billions of 1975 S
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
2150
2251
2372
2485
2641
2781
2931
3039
3213
3385
3501
3635
3830
4061
4094
4080
4284
4449
4618
4776
4835
.38
.52
.62
.74
.22
.35
.17
.67
.59
.03
.64
.82
.22
.50
.72
.64
.42
.49
.97
.08
1 T
Population
) (1000s)
656
664
672
679
686
693
699
706
714
723
730
737
744
750
756
762
768
774
780
NA
NA
,444
,384
,171
,749
,837
,529
,995
,919
,281
,180
,617
,699
,409
,860
,724
,476
,355
,275
,643
GDP/Population
(1975 S/person)
of GDP/Population
C°.)
NA
NA
3614.
3741.
3929.
4091.
4267.
4382.
4590.
4788.
4902.
5027.
5242.
5505.
5500.
5434.
5661.
5835.
6011.
6168.
6193.
3
4
4
7
6
9
9
4
3
5
4
6
6
6
8
6
5
4
9
3
5
4
4
2
4
4
2
2
4
5
-0
-1
4
3
3
9
0
NA
NA
--
. 5
.0
.1
.3
.7
.7
.3
.3
.6
.3
.0
.1
9
. L.
.2
.1
.0
.6
.4
a/ See attached footnotes.
NA = not available.
-------
A-18
TABLE A-7: OECD ECONOMIC DATA
Description: Table A-7 presents time series data on OECD GDP (measured at
1975 exchange rates and in billions of constant 1975 U.S.
dollars), population, and GDP/population. Data are only
available for the period 1960 to 1980.
Footnotes: a) OECD. National Accounts: Volume 1, Main Aggregates
1951-1980. 1982.
-------
A-19
TABLE A-8
PROJECTIONS OF GNP AND POPULATION
RELATIVE TO 1975 a/
OECD REGIONS
UNITED
POPULA
1.0000
1.1890
1.3170
1.3470
1.3664
1.3664
EASTERN
AND
POPULA
1.0000
1.1960
1.3080
1.3510
1.3720
1.3720
STATES
GNP
1.000
1.877
3.021
4.505
6.537
9.484
EUROPE
USSR
GNP
1.000
1.815
2.943
4.442
6.445
9.351
CANADA AND
WESTERN EUROPE
POPULA GNP
1.0000 1.000
1.1760 1.978
1.3030 3.534
1.3650 5.709
1.3872 8.283
1.3872 12.019
CENTRALLY PLANNED AND
CENTRALLY
PLANNED ASIA
POPULA GNP
1.0000 1.000
1.3700 2.665
1.6450 5.468
1.7700 10.335
1.8081 14.996
1.8081 21.758
OECD PACIFIC
POPULA GNP
1.0000 1.000
1.2010 2.701
1.2790 5.432
1.3050 9.121
1.3176 13.234
1.3176 19.202
MIDEAST REGIONS
MIDDLE EAST
POPULA GNP
1.0000 1.000
1.8030 3.649
2.4480 8.959
2.8470 18.044
2.9595 26.181
2.9595 37,987
YEAR
1975
2000
2025
2050 b/
2075
2100
YEAR
1975
2000
2025
2050
2075
2100
DEVELOPING COUNTRY REGIONS
AFRICA
POPULA
1.0000
1.7450
2.3610
2.7580
2.8807
2.8807
GNP
1.000
3.069
7.327
13.806
20.032
29.065
CENTRALLY
LATIN AMERICA
POPULA GNP
1.0000 1.000
1.7280 3.475
2.2980 9.005
2.6310 16.395
2.7144 23.788
2.7144 34.515
SOUTH AND
EAST ASIA
POPULA GNP
1.0000 1.000
1.6850 3.079
2.2260 7.169
2.5560 13.502
2.6504 19.591
2.6504 28.425
YEAR
1975
2000
2025
2050
2075
2100
b/
b/
a/ See attached footnotes.
-------
A-20
TABLE A-8:
PROJECTIONS OF GNP AND POPULATION
RELATIVE TO 1975
Description:
Footnotes:
Table A-8 presents projections of GNP and population for the
nine regions of the world: 1) U.S.; 2) Canada and Western
Europe; 3) OECD Pacific; 4) Eastern Europe and USSR; 5)
Centrally Planned Asia; 6) the Middle East; 7) Africa; 8) Latin
America; and 9) South and East Asia.
a) Institute for Energy Analysis, Oak Ridge Associated
Universities. The values in the table are relative to
1975. For example, in the U.S., population is projected to
be 1.3664 times the 1975 population by 2100.
b) The post-2050 GNP annual growth rates were assumed to be
1.5 percent in all parts of the world.
-------
ANNEX B
REGRESSION ANALYSIS RESULTS
The equations used to project scenarios of CFC use were estimated using
least squares regression. The equations were evaluated using t-statistics and
adjusted R-squared values (R2). Most importantly, the variable coefficients
were examined to see whether their signs and sizes conformed to theoretical
and intuitive expectations.
The regression equations estimated are presented in Exhibit B-l. The
t-statistics were used to determine whether the coefficients are statistically
significant at the 5% level. (The t-statistics are shown in parentheses below
their corresponding coefficients.) The R2, also presented in Exhibit B-l,
measures the proportion of variation in the dependent variable explained by
the right hand side variable(s), while controlling for the numbers of
explanantory variables.
In general, the regression equations explain a large portion of the
variation in the dependent variables. For example, in equation 1, the
variation in U.S. GNP per person explains 90 percent of the variation in the
U.S. use per person of CFC-11 in aerosol applications. The coefficient on the
U.S. GNP per person variable implies that a one unit increase in this quantity
will correspond to a 1.2E-07 unit increase in CFC use per person (in millions
of kilograms).l
1 The format "1.2E-07" shown in Exhibit B-l means 1.2 times 10 raised to
the -07 power.
-------
EXHIBIr B-1
FSllMAltl) RELATIONSHIPS BFTWFEN CFC USE
AND FXPLANA10RY VARIABLES J./
I)op_erideia_Va rj.a^tile
I. U.S. CFC-1 I aerosol use/person
2. U.S. CFC-II nonaerosol use/person
3. U.S. CFC-I2 aerosol use/person
4. U.S. CFC-I2 nonaerosol use/person
5. U.S. CFC-22 use/person
6. U.S. CFC-II3 use/person
7. OECD-U.S. CFC-II aerosol use/person
8. OECD-U.S. CFC-II nonaerosol use/person
9. OECD-U.S. CFC-12 aerosol use/person
10. OECD-U.S. CFC-12 nonaerosol use/person
Exf) lanatory Variables and Coefficients
-3.IE-OU + I.2E-07 x U.S. GNP/POP
(-6.il) ( I I .76)
-I.7E-05 + 8.5E-08 x U.S. GNP/POP
(-0.2) (9.0)
+ -I.7E-06 x Price of CFC-II
(-6.U)
7./E-05 + I.2F-07 x U.S. GNP/POP
(O.U) (»4.2)
+ -2.0E-06 Price of CFC-12
(-3.0)
-7.'4F-Oi4 -i- 2.0E-07 x U.S. GNP/POP
(-'1.8) (7.5)
-6.2F-OU + I.6E-07 x U.S. GNP/POP
(-5.0) (7.3)
-3.7F-OU + 8.9E-08 x U.S. GNP/POP
(-10.3) (12.8)
-2.9F-OU + I.IF-07 x OECD-U.S. GNP/POP
(-23.6) (36.0)
-2.9E-OU + 9.2F-08 x OECD-U.S. GNP/POP
(-12.5) (16.8)
-2.6F-OU + I.IE-07 x OECD-U.S. GNP/POP
(-20.I) (31.U)
-6.0F-06 +A.3F-03^< OECD-U.S. GNP/POP
(-0.2) \HJL2±^
+ -5.07E-07 Price of CFC-12
(-U.7)
Adjusted R-Squared
0.90
0.97
0.98
0.89
0.88
0.89
0.99
0.9U
0.99
0.98
J/ CFC use/person in thousands of kilograms per person. U.S. GNP/POP is in 1972 dollars per person.
OFCD-U.S. GNP/POP i" 1975 dollars per person. Prices of CFC-II and CFC-12 expressed as an index with 1972 = 100.
-------
ANNEX C
THE RELATIONSHIPS AMONG CFC USE,
CFC EMISSIONS, AND BANKED CFCs
Annual CFC use and emissions are not equivalent. Some portion of the use
may be released immediately (e.g., emissions from solvent use). However, a
large portion may remain in the product for many years (e.g., the refrigerant
trapped in the appliance). The CFCs remaining in products are referred to as
"banked." Banked CFCs are emitted slowly over time (e.g., when a refrigerator
compressor is repaired or disposed of, the CFCs in the refrigerator may be
released). Exhibit C-l displays the manner in which CFC use, emissions, and
the CFC bank interact and are treated in this analysis.
In year 1 of the analysis (e.g., 1985) there is an existing amount of
banked CFCs from past uses of products containing these chemicals. These CFCs
contribute to emissions in year 1. Also in year 1, there is some level of CFC
use. Some portion of these CFCs are emitted promptly, and become part of the
year 1 emissions. The remaining portion is banked, and becomes part of the
bank in year 2. These banked CFCs are then slowly emitted from year 2
onward. This procedure repeats year after year.
From the schematic in Exhibit C-l it is clear that the following three
components must be specified in order to project scenarios of emissions:
• the future rate of CFC use;
• current reservoir of banked CFCs; and
• the rate of prompt emissions directly from use (e.g.,
from CFC use as aerosol propellants) and the rate of
emissions from the banked CFCs (e.g., the rate of
releases of CFCs trapped in appliances).
The future rates of use are developed as scenarios. Estimates of the
current CFC bank and the rates of emissions from use and the bank were
developed. The total world quantity of currently banked CFC-11 and CFC-12 was
estimated as 768.9 and 880.7 million kilograms respectively.1 No estimate
was available for the amount of CFC-22 banked. Because CFC-113 is only used
in a prompt emitting applications, there is no CFC-113 bank.
The release rates from the existing CFC-11 and CFC-12 banks are displayed
in Exhibit C-2. For example, the exhibit shows that over the next ten years,
27.5 percent of the banked CFC-11 is expected to be released. Of note is that
a fairly large fraction of CFC-11 (22.8 percent) is expected to take a very
long time to be released. These CFCs are banked in insulating foams.
"Production, Sales ...", CMA, ojg. cit. These data are for 1982.
-------
C-2
EXHIBIT C-l
SCHEMATIC OF CFC USE, EMISSIONS, AND BANKS
EMISSIONS:
YEAR 1
| BANK:
1 YEAR 1
USE:
YEAR 2
! USE:
! YEAR 3
EMISSIONS:
YEAR 2
->• ! EMISSIONS: j
I YEAR 3 ;
BANK:
YEAR 2
-------
C-3
EXHIBIT C-2
EMISSIONS FROM EXISTING BANKS
(percentage)
Years from Now CFC-11 CFC-12
1-10 27.5 85.0
11-20 27.0 6.3
21-30 6.2 7.5
31-40 3.9 0.1
41-50 3.8 0.1
51-60 3.3 0.1
61-70 2.8 0.1
71-80 2.7 0.1
81-150 22.8 0.7
Total 100.0 100.0
Source: See text.
Exhibit C-3 shows the rate of release after the use of CFCs in non-aerosol
applications.2 For example, 66.4 percent of the use of CFC-11 in
non-aerosol applications is estimated to be released promptly, i.e., in the
year of use (the exhibit shows this 66.4 percent being released zero years
after use). Again, CFC-11 shows a substantial fraction being banked over a
long period of time.
These release rates were derived by: (1) identifying the mix of
applications of each CFC (i.e., the fraction going to air conditioners, foams,
refrigerators, etc.); (2) identifying the rate of emissions from each of these
applications; and (3) weighting the rates of emissions by the mix of the
applications to estimate release rates for the CFC type. The data describing
the rates of release by application and the mix of applications are presented
in Exhibit C-4. These release rate estimates were developed from Environment
Committee Report on Chlorofluorocarbons by OECD (1982). The mix of uses is
for the U.S. in 1976 as reported in Economic Implications of Regulating
Chlorofluorocarbon Emissions from Nonaerosol Applications, by The Rand
Corporation (June 1980). The emissions rates reported in Exhibits 3 and 4
were used for all parts of the world, throughout the entire period examined.
2 The release rate for the aerosol applications of CFC-11 and CFC-12 as
well as all the use of CFC-113 is assumed to be 100 percent in the year of use.
-------
C-4
EXHIBIT C-3
EMISSIONS FROM USE
(percentage)
Years After
Use
0
1
11
21
31
41
51
61
-10
-20
-30
-40
-50
-60
-70
71-80
81
-150
CFC-11
Non-Aerosol
66,
7,
7 ,
3,
2.
2,
2.
1
1
4
,4
.6
.2
.6
.5
,4
,1
.8
.7
.7
CFC-
12
Non-Aerosol
45
46
3
4
0
0
0
0
0
0
.0
.5
.7
.4
.1
.1
.1
.1
.0
.0
CFC-22
Non-Aerosol
45,
23,
23,
7
0
0
0
0
0
0
.0
.9
.4
.7
.0
.0
.0
.0
.0
.0
Total 100.0 100.0 100.0
Source: See text.
-------
RXHIBIT C-4
PROMPT VS. BANKED EMISSION DATA BY END USE
CPC
Product
CPC- 11
CFC-12
CFC-22
CPC- 113
Principal End Use
and Percent of
Product Use
Flexible Foam: 34%
Rigid Foams
Urethane: 35%
Nonurethane: 2%
Chillers: 8%
Other: 20%
Rigid Poams
Urethane: 1%
Nonurethane: 11%
MACS2: 48%
Chillers: 3%
H one Re f r I g . &
Freezer: 3%
Retail Pood
Refrlg. : 6%
Other: 28%
Chillers: 3%
Retail Pood
Refrlg.: 1%
Other: 96%
Solvents: 100%
Emissions
%
Prompt
100%
26.5
100.0
2.0
100. 0
26.5
100.0
11.0
2.0
7.0
3.0
100.0
2.0
3.0
100.0
100. 0
%
Banked
0%
73.5
0
98.0
0
73.5
n
89.0
98.0
93.0
97.0
0
98.0
97.0
0
0
Ypar of Release of Banked CFCs After Initial Use qreater
Life of 31- 41- 51- 61- 71- 81- 91- than
End use 1-10 11-20 21-30 40 50 t>0 70 80 90 100 100
1/2 life of 11% 10% 9% 7% 7% 6% 5% 5% 4% 4% 32%
emissions
= 60 year s
20 years 47.5 (46.5)1 47.5 (46.5) 5.0
1/2 life of 11 10 97765544 32
emissions
= 60 years
7 years 100 ( 89)
20 years 47.5 (46.5) 47.5 (46.5) 5.0
20 years 16.0 (14.5) 16.0 (14.5) 69.0 (64)
20 years 31.0 (30) 31.0 (30) 38.0 (37)
—
20 years 47.5 (46.5) 47.5 (46.5) 5.0
20 years 31.0 (30) 31.0 (30) 38.0 (37)
— —
—
1 Nuntoers in parentheses refer to the rate of release of total CPCs emitted (i.e. Prompt and hanked emissions).
2 Mobile Air Conditioners.
Source: ICF analysis of data in: Environment Committee Report on Chlorofluorocarbons, OECD, 1982, and Economic Implications of Regulating
Chlorofluorocarbon Emissions from Non-Aerosol Applications, The Rand Corporation, June 1980.
-------
ANNEX D
SCENARIOS BY TYPE OF CFC
This annex presents a series of tables that describe the potential future
use and emission of CFC-11, CFC-12, CFC-22, and CFC-113 in five scenarios:
Limits to Growth; Low; Medium; High; and No Limits to Growth. These estimates
are based on the assumption that current aerosol use restrictions remain in
force.
For each CFC type in each scenario the following information is provided
for the world total:
• annual use and the annual percent growth in use;
• total annual emissions;
• portions of emissions that come:
directly from use
from the banked CFCs created due to projected use
(referred to as the "simulated use bank" and
labelled "UBANK")
from the banked CFCs that already exist today
(referred to as the "other bank" and labelled
"OBANK");
• total banked CFCs, labelled "TBANK;"
• banked CFCs from simulated future use, labelled
"UBANK;" and
• banked CFCs that already exist today, labelled "OBANK."
-------
RESULTS FOR LIMITS TO GROWTH SCENARIO
-------
WOR10 IOTAI: ClC-11
COMI'ONI Nl S I NCI IJDI I) :
UNI III) SIAIIS NUN-AI KOSOI I)S( 01 CIC-11: IINIAK MODI I 2
ON I IN) SIAIIS Al KOSOI USI Of CIC-11: I I Nl AR MODI I + AlROSOI RE SI = I). 5
IIC NON-AI ROSOI USC Ol CfC-11: I I Nl AR MODll ?
IIC Al ROSOI USE 01 CSC-ll: MNIAR MODI L 2 + UIKR1NI Al ROSOI S RE SIR.
OI Cl) - US - I I C NON-AI ROSOI USI 01 CIC-11: MNIAR MODll 2
01 Cl) - US - I 1C Al ROSOt USI Ol CIC-11: I INI AH MODll 2
NON-OI CD NON-AI KOSOI USI 01 CIC-11: I OG I I / I'OI'UI A I I ON MODE I
NON-01 CO AIKOSOI USI Ol CIC-11: I 00 I I / I'OI'UI A I I ON MODll
EMISSION EKOM WORI D HANK 01 CEC-11: BISI I SI I MA IE - NO KISIRICIIONS
(All I SI IMAMS IN Ml I I IONS 01 KIIOGRAMS)
% ANNUAI
GROW III
1985
1991)
1995
2IIIIO
'Iin5
>(} 10
•015
'O2O
'0. '5
••() 11)
'II 15
>0'll)
'O'l5
2(151)
MI55
2060
2065
'070
'0/5
O
3
3
£>
1
1
1
1
1
1
1
2
2
2
1
1
1
1
1
0
3
0
(1
}{
8
9
9
9
9
9
0
0
0
•j
6
6
6
5
U S
Cl.'
'11)1,
'1 /'I
V4(>
598
655
/I /
/H /
865
9'lH
III'IO
1 1'|5
126 i
1 196
15D5
1625
1 757
1 899
2052
1
3
8
6
1
2
1
9
>)
'l
0
8
3
?
3
14
5
0
5
2
EMI SSIONS
IOIAI
29'l
351
'11 2
'1/9
517
5/2
631
696
7 /()
8 '1 8
935
1013
1 1 '1 3
1267
1.1/5
1'|92
1620
1 756
1902
d
/
9
1
0
3
2
9
0
t)
?
J
')
9
I)
2
1
/
6
IMI SSIONS
1 ROM U'
2/3
322
3/ 5
'|27
'168
513
562
618
680
/'I6
82 1
906
1003
1112
1203
1303
I'll 3
1533
1661
MISSIONS 1
MISSIONS
il (ROM UliANK (ROM OI1ANK
5
(1
1
/
3
n
6
0
•t
1
1
6
0
;j
0
'i
6
2
6
O
8
18
10
'I'l
5'l
65
75
86
98
1 10
1 2 'I
1 18
153
169
186
20'4
221
238
0
•>
8
6
0
5
(>
9
8
5
9
1
'I
6
9
/
0
0
6
21 .
21 .
20.
20.
'1.
l\ .
3.
3.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
1
1
8
8
8
8
0
0
9
9
5
5
2
2
1
1
5
5
5
IBANK
795
1056
1352
16/7
2039
?'l'l9
28/3
3318
3/85
'12/5
'1791
5338
5920
65'l't
7192
7852
8530
9232
9966
lj
8
3
5
2
6
3
2
2
6
7
5
9
9
3
8
3
2
5
UliANK
68
'l 3 5
836
1265
1698
2133
25/7
3037
3518
'1023
'4553
5113
5 70 7
63'l2
7000
76/1
8160
907I|
9821
8
9
14
3
9
1
1
0
9
8
8
3
6
'I
1^
3
0
U
2
OBANK
726.
620.
515.
1412.
3140.
316.
296.
281 .
266.
251.
237.
225.
213.
202.
191 .
181 .
170.
157.
1U5.
6
9
9
1
3
5
2
2
3
7
9
2
3
5
9
5
3
8
3
ANNUAI AVI RAGE GKOW1II RAIIS:
2.011 2.010
2.09
-------
WORl0 IOFAI : CFC-12
COMI'ONI NIS I NCI UOI I) :
UN I I CD SI AllS NON-AIROSOt USE Of CIC-12: IINIAK MOOT I 2
UNITED SIAIIS Al ROSOl US( OK CIC-12: IINIAR MODI 1. + Al ROSl RFS1 - '(. 5
tlC NON-ALROSOL USF 01 CIC-1?: I INtAR MODEL 2
TIC AEROSOL USE OE CFO-12: IINIAR MODEL 2 + CUKRtNl Al ROSOL RESER.
OICD - US - EEC NON-AI ROSOl USE OF CFC-1?: I IMIAR MODI I. 2
01 CO - US - EEC AEROSOl USE OE CEC-12: I INEAK MODE I 2
NON-OECD NON-AEROSOL USE Ol CIC-12: I 00 I I / POI'UI Al I ON MODEL
NON-OE CD Al ROSOl US! Of CEC-12: I 00 I 1 / POI'Ul A I I ON MODI I
EMISSIONS FROM WOULD HANK OF CEC-12: EHS1 ESIIMAIE - NO RESIRICIIONS
(AIL ESI I MATES IN MILLIONS OF KILOGRAMS)
ANNUAl
GROW III
0.0
2.9
2.7
2.5
.6
.6
.6
.6
.6
. 5
. 5
.5
. 5
.0
. 0
.0
.0
.0
u s
'160
535
61 5
696
75')
81 7
885
959
1039
1 120
1208
1 !03
11)06
1516
1681
1860
195't
I
2
1
5
2
2
0
1
6
6
6
1
2
3
I)
EMISSIONS EMISSIONS (MISSIONS (MISSIONS
10 (Al (ROM USl FROM UHANK E ROM OHANK
M09.9 335. 1 0.0 7').9
56/!/ 1)36.9 125.3 5.5
650.? ')92. 7 151.9 5.5
720
790
855
929
1007
1087
11 73
1266
1365
l')/3
1558
16') 3
1 730
1821
1911)
.2
.0
.8
.0
.8
.9
. 7
.2
.8
.2
_ ?
.8
.8
. 1
.2
535
582
633
688
7'l8
809
8/1)
9')5
1021
1 102
1161
1223
1287
1353
1U22
7
>j
3
6
6
;>
'j
1
1
9
9
l)
l|
/
2
1 //
200
222
2')0
259
2/8
299
321
3')')
3/0
396
<420
'l')3
1)67
1.91
9
9
')
J
1
6
1
0
/
2
2
i
3
3
9
6.6
6.6
0.
0.
0.
0.
0.
0. 1
0. 1
0. 1
0. 1
I). 1
0. 1
0. 1
0. 1
IBANK
856. 1
1016.6
1126.6
1357.9
1 552 .
1 700.
1836.
1985.
2138.
2299.
2'469.
2651.
28'! /.
305 /.
3259.
3')') 9 .
36MO.
3835.
14035.
14
5
7
2
7
3
5
6
1
2
5
5
1)
i)
1
UBANK
125.2
659.9
1005.6
1 26
-------
WORI U IOFAI : Cf O22
COMI'OW NtS INCIUDM) :
UNI IIU SIA1IS 101AL USL Of GIC-22: I INI AH MODI I 2
NON-US 01 CD USt 01 CIC-22: IINEAK MODI I 2 DtKIVH) I I
'(
2
'1
0
'£
1
O
8
2
0
O
?
6
MISSIONS f
MISSIONS EMISSIONS
1 HOM USI fKOM UHANK f ROM OBANK
33
'|2
01
60
6'l
67
71
76'
80
86
91
9 7
101
10
16
22
29
30
l'(2
5
i
'I
6
1
9
<•)
;>
9
1
6
•j
9
1
1
9
2
/
3
O
9
22
16
0 1
63
73
79
80
91
96
102
109
1 16
123
1 M
1 58
1 '1 6
1 O'l .
O
8
0
6
O
6
't
9
0
0
0
6
1
1
0
1
«
')
!
0.
0.
O.
0.
0.
o.
0.
0.
0.
0.
o.
0.
0.
0.
0.
0.
0.
0.
0.
0
0
0
0
0
o
0
0
0
o
0
0
0
0
0
o
0
0
0
IBANK
l|0.
2'(8.
'lOO .
600 .
801 .
909 .
990.
1008.
1 1 2'l .
1 1 9'l .
1269.
I3'|9.
Hi 36.
1 029 .
1620.
1 720.
1816.
1912.
2010.
7
'j
fj
3
l|
')
1
l\
9
2
0
7
'j
^
1
6
2
6
3
UBANk
'10
2'(8
'( 0 0
600
801
909
990
1008
1 1 2'l
1 19'l
1 269
1.3'l9
Hi 3 6
1029
1620
1 720
1816
1912.
2010.
7
0
•y
3
'1
0
7
il
9
2
0
7
0
0
1
6
2
6
3
OBANK
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
. 0
.0
.0
.0
ANNUAL AVI RAC[ (iKOWIII KAIIS:
1.6514 1.628
0.0
0.0
-------
WORI D 101AL: CFC-113
COMPONI NIS I NCI UDI I) :
UNI II I) STA1ES FOIAl US! Ol CIC-113: IINIAK MODI I ?
NON-US 01 CD USf 01 CFC-113: MNFAR MODH 2 DtKIVIl) I KOM U.S. USE
NON-OICO NON-AEKOSOI. USE. OF CtC-IU: LOG I'l /I'OI'lll A? I ON MOOtl
( Al L LSII MATES IN Ml I I IONS OF K11OGRAMS)
ANNUAI
GROW 1 M
0.0
'1.5
3. /
3.2
. 1
. 1
. 1
. 1
! i
.0
0.9
0.9
0.9
0.8
u s
1?.
92
1 12
13?
139
I'l 6
1'j'l
163
1 1?
182
191
20'4
216
??tt
?l|0
251
263
275
28 /
LMISSIONS EMISSIONS (MISSIONS EMISSIONS
E 101 Al FKOM US! 1 ROM UBANK FROM OBANK
.2
. 1
. 1
.3
.3
.9
.8
.3
. 3
.3
. 0
.2
.2
.9
.3
.9
.6
.5
.5
I?.
9?.
1 12.
13?.
139.
1«l6.
15'l.
163.
1 12.
18?.
193.
20'» .
?16.
2?8.
?'|U.
251.
?63.
2/5.
287.
2
1
1
3
3
9
8
3
3
3
0
2
2
9
3
9
6
5
5
/?
9?
1 1?
13?
139
I'l6
1'j'l
163
172
182
193
20'»
216
2?8
?'lO
?")!
263
?75
287
2
1
1
1
3
9
8
3
3
3
0
2
2
9
{
9
6
')
•>
0.
0.
().
0.
0.
0.
0.
0.
0.
0.
o.
0.
0.
0.
0.
o.
0.
0.
0.
0
0
1)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
o
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
TBANK
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
UBANK
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
(I
0
0
0
.0
.0
.0
.0
. 0
.0
.0
. 0
.0
.0
. 0
.0
.0
.0
.0
.0
.0
.0
.0
OBANK
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1985
1990
1995
21)00
2O05
2010
2015
2020
2025
2030
2(1 !5
20'U)
20'l5
20SO
20'; 5
2060
2065
211/0
?0/5
ANNUAL AVI RAGE GROWTH RAMS:
1 . 554 1 . 51
1 . 5't 7
0.0
0.0
0.0
0.0
0.0
-------
RESULTS FOR LOW SCENARIO
-------
WORLD TOTAL: CFC-11
COMPONINIS INCLUDED :
UNITED STAIES NON-AEROSOL USE OF CFC-11: LINIAR MOOR 2
UNIIH) SIATES AEROSOL USE OE CFC-11: LINIAR MODFI + AlROSOl RES I = 4.5
EEC NON-AEROSOL USE OE CEC-11: LINEAR MODEL 2
EEC AEROSOL USE OF CFC-11: LINEAR MODEL 2 + CUKRFNI Al ROSOLS RESTR.
OECD - US - EEC NON-AIROSOl USE OE CFC-11: LINIAR MODI L 2
OECO - US - EEC Al ROSOl USL OF CIC-11: MNEAR MODF I 2
NON-OECD NON-AEROSOL USE 01 CFC-11: I OCI I/POPUI AT I ON MODFI
NON-OFCD AEROSOL USE 01 CEC-11: LOCI f/POI'ULAl I ON MODII
EMISSION FROM WORlD BANK Of CEC-11: BEST ESTIMATE - NO RFSIRIC1IONS
(AIL ESTIMATES IN MILLIONS OF KILOGRAMS)
% ANNUAL
GROW 1 H
1985
1 990
1995
2000
2O10
2015
2020
2025
2O 30
2O'|0
2O'45
2O 50
2055
2060
2O65
20/0
20/5
0
3
3
3
2
2
2
2
2
2
2
2
2
2
2
2
2
1
. 0
. 3
.8
.5
.0
.9
.8
. 7
. 7
.2
.2
.2
.2
. 1
.14
. 3
. 1
.0
.8
U S
'10(1
'199
60 /
8<4'l
9/6
1 122
1 '4 / 1
16'4?
1832
20'43
2276
2533
2855
3199
356 i
3939
U322
E
j
.2
. 0
.5
.8
.3
.5
.3
.'4
.9
.7
.2
!2
.?.
. 7
.0
.5
.7
EMISSIONS
TOTAL
336
1421
515
715
83»4
965
1113
1281
lt'42
1619
1816
203'4
2275
2568
288'4
3221
35/1
3931
.5
. 0
.'4
.2
.9
.5
.6
.7
.14
.0
.5
.7
. 7
.0
.4
.9
. 7
.9
.0
EMISSIONS t
MISSIONS EMISSIONS
(ROM USf (ROM UBANK FROM OBANK
315
389
'4/0
559
651
/52
866
9 9 '4
1 I'lO
12/7
1'429
1600
1/89
1999
2259
2539
283'4
31 '41
3<453
. 3
. 1
. ;>
.9
.3
. 7
2
. >)
.H
. ?
]
. o
/
1
.9
.3
/
. 3
.5
0
10
2'4
'lO
59
77
96
1 16
138
161
187
2I<4
2'42
2/3
306
3'(3
38't
'428
t/5
. 0
.8
.'4
.6
.8
. 1
. J
.2
. 1
.9
.3
. y
. 8
.2
.5
.5
.5
. 1
.0
21
21
20
20
*4
It
3
3
2
2
2
2
2
2
2
2
2
2
2
. 1
. 1
.8
.8
.8
.8
.0
. 0
.9
.9
.5
.5
.2
.2
. 1
. 1
.5
.5
.5
TBANK
812.
11/5.
1606.
2103.
2682.
3365.
'4117.
'49'49.
5863.
68<46.
7886.
8992.
10169.
1 1'427.
1280'4.
1'4322.
1 59 76 .
1/762.
19673.
0
5
8
2
9
7
8
0
7
2
7
2
7
5
1
3
7
7
9
UBANK
85.14
55'4.6
1090.9
1691 . 1
23'l2.6
30149.2
3821.6
'4667.9
5597.14
659U.5
/6'48.8
8/6/.0
9956. tl
1 122U.9
12612.2
1 '4 1'(0.8
15806. «t
1/6014.9
19528.7
ANNUAL AVERAGE GROWTH KATFS:
2.680 2.6/8
OBANK
726.6
620.9
515.9
3<4o!3
316.5
296.2
281 .2
266.3
251 .7
237.9
225.2
213.3
202.
191
181
170.
157.8
1<45.3
.5
.9
.5
.3
2.769
2.695
0.0
-2.3'42
3.605
6.222
-1.773
-------
WORLD TOTAL: CFC-12
COMPONENTS INCIUDfl) :
UNI HI) STAILS NON-AtROSOl USF OT CfC-1?: I INIAR MODf I 2
UNIIEl) SIAIES AEROSOl USE OJ CEC-12: I INIAR MODEL + ALROSL REST = 4.5
EEC NON-AEROSOL USE Of CFC-12: IINEAR MODEL 2
EEC AEROSOL USE OF CFC-12: LINEAR MODEL 2 + CURRENI AEROSOL RESTR.
OECD - US - EEC NON-AIROSOL USE OF CFC-12: I INIAR MODI L 2
OfCO - US - EEC AEROSOL USl OF CFC-12: I INfAK MOOfl ?
NON-OFCD NON-AEROSOL USE 01 CfC-12: LOG I 1/POI'UI AT I ON MODEl
NON-OECD AEROSOL USE OT CIC-12: I OG I T/POI'W Al I ON MOON.
EMISSIONS FROM WOULD BANK OF CfC-12: BESF ESIIMAIE - NO RESTRICTIONS
(AIL ESTIMATES IN MILLIONS OF KILOGRAMS)
ANNUAI
GROWTH
0.0
3.9
3.6
3.3
2. 7
2.6
2.5
2.5
2.4
1.8
1 .8
.7
. 7
.6
.9
. 7
.6
.5
1.4
USE
545.3
666 . 5
/99.8
946 . 5
1083.3
123J.6
1399.0
1581 . 1
1/81.6
1 952 . 1
2133.4
2325.8
2529.0
2/42.8
3D13.0
3292.5
3579.7
3872.5
4168.5
EMISSIONS
TOTAL
457.2
612.0
724.8
866 . 9
1010.0
1163.5
1323.7
1500.9
1695
1871
2052
2242
2442
2652
2900
3167
3447
3734
4026
.5
.0
.6
.2
.3
. 7
.7
.0
.3
.6
.8
EMISSIONS (MISSIONS EMISSIONS
TROM USl FROM UBANK FROM OBANK
382.4 0.0 74.9
461.2 75.9 74.9
548.9 170.3 5.5
646.7 214.6 5.5
742.0 261.4 6.6
847.8 309. 1 6.6
965.0 358.6 0.1
1095.0 405.8 0.1
1238
1362
1493
1633
1/81
1937
2125
2319
2519
2723
2929
y
1
b
•t
6
6
1
4
?
O
1
456
508
558
608
660
715
7/5
847
928
1011
1097
.5
.8
.9
.5
.6
.0
.5
.5
.0
.5
.6
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
1
1
1
1
1
1
1
1
1
1
1
IBANK
893.9
1239.5
1510.5
1899.6
2278.3
2633.8
2993.6
3384.3
3802.
4216.
4619.
5031 .
5458.
5902.
6423.
7028.
76/7.
8356.
9058.
6
6
4
6
9
7
4
5
4
7
5
UBANK
162.9
882.8
1389.5
1806.3
2214.9
2603.4
2983.2
3374.4
3/93.1
4207.5
4610.7
5023.4
5451.1
5895.4
6416.5
7022.0
76/1.4
8351.2
9053.4
1985
1990
1995
2000
2005
2010
2015
2020
2025
2030
2055
2040
2045
2050
2055
2060
2065
2070
2075
ANNUAI AVERAGE GROWTH RATES:
2.288 2.286
2.447
2.288
0.0
-7.221
2.607
4.565
OBANK
731.0
356.7
121 .0
93.3
63.4
30.4
10.4
10.0
9.5
9. 1
8.6
8.2
7.7
7.3
6.9
6.4
6.0
5.5
5.1
-5.366
-------
WORI0 TOTAl: CFC-22
COMI'ONI N ! S I NCI UOEI) :
UNIIED STAILS TOTAl USI Of CIC-22: IINIAK MOON 2
NON-US OECO USE 01 CFC-22: UNLAR MODtl 2 I)E in VI D I KOM U.S. USE
NON-OECD NON-AEKOSOI USE Of CFC-22: LOGI T/POI'Ul A[ ION MOOt L
(AIL ESTIMATES IN Ml I LIONS OF KILOGRAMS)
1985
1 990
1995
2000
2005
2010
2015
2020
2025
2030
20.55
20'IO
20')0
2055
2060
2065
20/0
20/5
ANNUAL
GROW I II
0.0
5. 1
'1.3
3.8
2.U
2.3
2.2
2. 1
2.0
.5
.3
. 3
.9
.8
. /
.6
.5
EMISSIONS EMISSIONS EMISSIONS EMISSIONS
USE
109.5
1 l| 3 . 7
180.2
218.9
21)6.9
276. 7
308.6
5142.8
579.5
'109 . 1
i|39. 7
'(71.5
50'l . 5
53H.6
59'i . 9
652 . 3
no. /
769 . 8
829 . 3
TOIAl
'49
79
11<4
155
195
229
265
298
332
36')
395
1)28
'460
»493
537
585
637
691
7t«9
. 3
.14
. 7
.5
.5
.6
. 3
. 5
.8
.2
.9
.0
.5
.5
.8
. 7
. 1
.9
.2
H
-------
WORl 0 IOIAI : CfC-1 1 5
COMI'ONI NIS I NCI Ul)[ I) :
UNIHDSIAIIS I01AI USE 0( CIC-IU: I I Nl AH MODI I 2
NON-US 01 CO USt 01 CH> 113: LINIAR MODI I 7 UIKIVIO H(OM U S
NON-OhCD NON-AF.KOSOL USE OF CfC-113: LOG II/POI'UI AT I ON MODf L
(All t'SIIMAIES IN Ml I I IONS 01 KI IOGRAMS)
% ANNUAI
GKOWIII
1985 0.0
1 990 5 . 5
1995 '1.6
2OOO 3.9
?005 2. '4
2010 2.3
2(115 2.2
21)20 2. 1
2(>;"> 2.0
20 10
20 i 5
20'|0
2O'|5
2(l')0
2 (155
2060
2065
20/0
20/5
.5
.'1
.3
. 3
. 2
.9
.8
. /
.6
• 'I
U S
9 »
125
150
33 /
363
)90
'll/
'i'l5
'!/ i
5? I
57 i
6 2 'l
6/6
72H
F
. 1
9
0
8
3
6
8
6
1
3
2
9
'1
9
9
5
"4
FMI SSIONS EMISSIONS (MISSIONS EMISSIONS
IOIAI (ROM USt (ROM UBANK FROM OBANK
93.2 93.2 0.0 0.0
125. 1 125. 1 0.0 0.0
158.9 158.9 0.0 O.O
19'l. / 19'l. / O.o O.o
220.0 220.0 0.0 0.0
2'46.8 2'l6.H O.I) 0.0
2/5.3 2 1'j. j O.O 0.0
305 . 6 305 . 6 0 o o 0
337
363
390
'41 /
'1 '4 5
*4/3
523
5/3
62'i
6/6
728
8
6
1
3
2
9
<4
9
9
5
*4
337
363
390
'117
*I'I5
'1/3
523
5/3
62'l
676
728
H
6
1
3
2
<)
'1
9
9
V
'4
0
0
0
0
O
0
0
0
0
0
0
1)
I)
1)
(1
II
I)
0
0
1)
0
0
0.
0.
0.
0.
o.
0.
0.
0.
0.
0.
0.
0
0
0
0
0
0
o
0
0
0
0
IBANK
0.0
0.0
o.o
o.o
0.0
0.0
0.0
0.0
o.
o.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0
0
0
0
0
0
0
0
0
0
0
UBANK
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0
0
0
0
0
0
0
0
0
0
0
OBANK
0.0
0.0
0.0
0.0
0.0
0.0
0.0
on
o
o
o
0
0
0
0
0
o
0
0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
ANNUAL AVI RAGt GKOVUII RAITS:
2. 319
2.311
2.311
2.311
0.0
0.0
0.0
0.0
0.0
-------
RESULTS FOR MEDIUM SCENARIO
-------
WORl0 TOIAL: CFC-11
COMPONINIS INCIUUM) :
UNI III) SI AHS NUN-AI ROSOL USI OF CIC-11: IINIAK MODI I 2
UNI III) SIAIbS AIROSOL USE 01 CIC-11: LINIAR MODI.L + AlKOSOL REST = 4.
(1C NON-AlROSOl USE 01 CIC-11: LINEAR MOLJEL 2
EEC AEROSOL USE 01 CEC-11: IINEAR HOUEl 2 + CIIKKENI AIROSOLS RES1R.
OECD - US - EtC NON-Al ROSOI USI OF CIC-11: IINIAK MODI I 2
OfCD - US - EEC AEROSUl USI OE CEC-11: LINIAIi MODEL ?
NON-OECD NON-AE ROSOI USE 01 CIC-11: L OG I 1 /TOI'lll Al I ON MODEL
NON-01 CO AEROSOL USt 01 CfC-11: LOG I I/POI'UI Al I ON MODI L
EMISSION EROM WORl D BANK 01 CFC-11: BEST ESTIMATE - NO RESTRICTIONS
(AIL ESI (MATES IN Ml I LIONS OF KILOGRAMS)
1985
19'JO
1995
200(1
2005
2010
2015
2020
2025
2030
20 i 5
2040
2045
2050
2055
2060
2065
2070
20/5
ANNIMl
GROW III
0.0
4. /
4.2
4 . 0
4. 1
3.9
3.8
3. /
3.6
3.7
3.5
3.2
3.0
2. /
2.6
2.2
1 .9
1 .6
1 .4
EMISSIONS
U S
408
519
643
/85
966
11 /'(
1416
1 700
2O34
24 5<>
293(1
3 '( 5 5
4022
4616
52 //
5920
6531
7105
7636
E
. l|
.0
.5
.4
.9
.8
.5
. 1
. /
.2
.6
.6
.3
. 3
.8
.2
.0
.6
.6
TOTAL
342.
436.
545 .
6/0.
816.
998.
1211.
1463.
1 /62.
2134.
2558 .
3U31 .
3546 .
4091.
468 / .
52/5.
5843 .
6382.
6889.
5
8
0
9
0
5
0
1
2
9
t|
4
4
0
9
9
7
1
9
EMISSIONS EMISSIONS EMISSIONS
1 ROM USI I ROM UUANK FROM OBANK
321
404
498^
607
74 7
909
1 099
1 324
1592
1928.
2309
2733.
3192
56 74
4202
4713
5198
5649
6067.
. 3
.5
.9
. s
. /i
.2
. 1
.2
. (I
, l|
.5
ll
'. 1
.9
.6
.9
. 2
.6
.2
0
1 1
25,
42,
63,
84 .
108,
135,
167,
203,
246
295,
351 ,
413,
483
559,
643,
730
820,
. 0
, 1
.4
.6
.8
.5
.9
.9
.2
.6
. \
.5
.5
.9
.2
.9
. 1
. 0
.2
21.
21 .
20.
20.
14 ,
4.
3.
3.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
1
1
8
8
8
8
0
0
9
9
5
5
2
2
1
1
5
5
5
TBANK
813.
1192.
1652.
2193.
2859.
3689.
4656.
57/6.
/U66.
8575.
10333.
12350.
14627.
1/156.
19980.
2309/.
26452.
29995.
33681.
7
8
5
0
0
5
1
9
9
3
8
7
2
4
/
6
8
8
7
UBANK
87,
571
1136.
1 /80.
2518,
33/3.
4359,
5495,
6800,
8323,
10095
12125
14413,
16953,
19/88
22916,
26282,
29838
33536
. 1
.9
6
8
.7
.0
.9
.7
.5
.5
.9
.5
.9
.9
. 7
. 1
.5
.0
.5
OBANK
726.6
620.9
515.9
412. 1
340.3
316.5
296.2
281 .2
266.3
251.7
23/.9
225.2
213.3
202.5
191.9
181.5
1 /O. 3
157.8
145.3
ANNUAL AVERAGE CROW1H RATES:
3.312
i.307
3.391
3.319
0.0
-2.342
4.224
6.838
-1.773
-------
WORID TOTAL: CFG-12
COMI'ONFNIS INCIUDII) :
UNIIFD SIAILS NON-Af ROSOI USE 01 CIC-1?: I I Nl Al< MODI I 2
UNIIfl) SIAIFS Al KOSOI USL Oh CFC-12: UNIAR MODI L + At ROSL RfSF = 4
FfC NON-AFROSOL USI OF CFC-12: UNFAR MODEL 2
HC A( ROSOI USE 01 CFC-12: I INFAR MODF I ? -I- CIJRKtNl At ROSOL RESFR
01 CD - US - ILC NON-At ROSOI USL Of CFC-12: I INIAR MODI L 2
01 CD - US - FFC At ROSOI USt OF CfC-12: IINtAK MODU 2
NON-OFCD NON-AFROSOI USL 01 CFC-12: I OC I I /POI'UI AT I ON MODF I
NON-OFCD AFKOSOI USF Of CTC-12: LOG II /CO I'Ul A I I ON MODI I
(MISSIONS FROM WORLD HANK OF CFC-12: FHSI LSMMAIE - NO RESTRICTIONS
(AIL LSdMATES IN MILLIONS OF KILOGRAMS)
1985
1990
1995
2000
2005
2(1 10
2015
2020
2025
20 id
20 i5
20'|0
21H45
2050
2055
2060
2065
2070
2075
ANNUAL AVIRAGF CROW!!! RA1IS:
ANNUAI
GROW III
0.0
'4.6
'1.2
(4.0
3.8
3.6
3^3
3. 1
3.0
2.8
2.5
2. 3
2. 1
2.0
1 .8
1 .6
1 .<4
1 .2
u s
558
/O'l
871
106'4
1287
15441
1831
2158
252')
29'l'l
3390
3858
'I3'40
'4828
53/2
5900
6'l()/
689?
735'!
E
. 7
.5
.6
. 3
.5
.9
.2
.6
.8
.8
.8
.'I
. 1
.0
.6
.14
.6
.2
.2
EMISSIONS
TOFAl
<467.9
61(3.9
788. '1
973.5
1189. 7
1137.5
1 7 1 U . 2
2028
2381
2 / 7*4
3202
3656
'4 1 26
'4606
5115
5628
613'4
6620
708 /
5
2
8
2
0
6
5
0
1
1
1
2
(MISSIONS (MISSIONS EMISSIONS
(ROM USI FROM UBANK FROM OBANK TBANK
393. 1 0.0 7'4.9 896. /
'•91.1 78.0 7(4.9 1266.2
606.0 176.8 5.5 157'4.(4
7*41.0 226.9 5.5 20 13. '4
899.8 283.2 6.6 2'i89.9
1083.0 3'4/.9 6.6 2998.8
1293.2 1420.9 0.1 3552.0
1532
1802
2100
2'419
2752
3095
3'l'4l
380'!
'H55
'('489
'1806
5106
6
0
9
3
6
1
o
7
1
5
6
5
1495
579
673
782
903
1031
1 165
1310
1'4/3
16 '4 '4
18 Hi
1980
.8
. 1
.8
.8
. '1
.3
.'I
. 3
. 0
.5
. 0
.6
0.1 14175
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
(1869
5671
6579
756(4
8611
970<4
10928
12268
13635
1(4998
163*43
.6
.2
. 1
.6
.7
.3
.3
.8
.9
.6
./4
UBANK
165.7
909.5
H453.14
1920. 1
2(426.5
2968. '4
35'll.6
'4165. 7
(4859.7
5662.0
6570.9
7556.5
8603.5
9697.0
10921 .9
1 2262 .*5
13629.6
H4992.9
16338.2
OBANK
731.0
356.7
121 .0
93.3
63.1
30.»4
10. U
10.0
9.5
9.1
8.6
8.2
7.7
7.3
6.9
6.14
6.0
5.5
5.1
2.910
2.905
3.066
2.890
0.0
-7.221
3.278
5.23(4
-5.366
-------
WORl D 101 Al : CIC-22
COMI'ONI NI S I NCI UDI I) :
UNI iiu siAirs IOIAI usi or cu;-22: IINIAH MODI i 2
NON-US OICD USF 01 cic-22: IINEAK MOOI i 2 DIRIVID i KOM u.s. usi
NON-OICD NON-AtKOSOI. USt 01 CrC-22: LOG I I/POI'UI A I I ON MODLL
( Al L I SI I MAI IS IN Ml I I IONS 01 KllOGRAMS)
% ANNUAL
GROW 1 M
1985
1900
199!)
20UO
2O05
?010
2015
2020
2025
;>o so
2iH5
2040
2045
2()')0
2055
2(160
2065
20/0
20/5
0,
5,
'1.
'1,
3.
3.
3
2
2
2
2
2
2
2
2
2
1
1
1
. 0
.6
.7
, 1
. '1
^>
. 0
.8
. /
.8
.6
.4
.2
. 0
.5
. 2
.9
. /
.-3
II S
1 10.
149.
190.
235.
28(1.
329.
!82.
'I'lO.
504.
58 i.
666.
752.
841.
931.
106(1.
1 I8/.
ini.
143 i.
r>5;>.
I
9
3
6
5
3
1
5
9
6
6
5
7
4
5
3
1
5
2
1
IMISSIONS 1
IOIAI
'19.
82.
120.
165.
216.
263.
315.
368.
425.
1)92.
565.
6'13.
725.
811.
918.
1030.
1 146.
1265.
1385.
9
2
6
/
1
3
4
2
'1
5
1
0
6
7
9
9
7
6
6
MISSIONS I
MISSIONS EMISSIONS
1 KOM US! I KOM IMANK f ROM OBANK
49.
61 .
05.
106.
126.
I'l8.
1 12.
198.
22 /.
262.
299 .
338.
378.
'119.
>>n.
53').
590.
6'4'l .
698.
9
'>
8
0
1
1
1
'I
1
6
9
/
6
O
1
•J
'1
9
1)
0.
15.
3«4.
59.
90.
115.
1')3.
169.
198.
229.
265.
30'),
3'l/,
392.
l)')l
')96
556
620
687
0
1
8
1
. 0
,2
, J
,8
lj
.9
.2
. 3
. 0
.5
. /
. /
.6
.6
. 1
0
0
0
0
o
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.0
. 0
. 0
. 0
. 0
.0
. 0
.0
.0
. 0
. 0
.0
. 0
. 0
. 0
.0
.0
.0
.0
IBANK
61
385
730
1081
1 ') 1 <)
1 710
2073
2425
2808
32')0
3727
1)259
'1827
5') 1 9
6086
68')0
76')9
8i48'4
9319
.0
.3
.6
. 3
. 7
. '1
.4
.14
.0
.2
.2
.8
.14
.2
. 1
.2
.U
.0
.5
UBANK
61.0
385. 3
730.6
1081.3
1 ') 1 ') . 7
1 7'40.l|
20/3.4
2425.4
28O8.0
3240.2
3727.2
4259.8
4827.4
5419.2
6086. 1
6840.2
7649.4
8484.0
9319.5
ANNUAL AVfRAGE GROWTH RAIIS:
2.981 2.975
3. 762
2.9/5
0.0
0.0
5. 747
5. 747
OBANK
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
-------
WORLO TOTAL: CFC-113
COMI'ONINIS I NCI 1)1)1 I) :
UNIITl) SIAIES IOIAI USf 01 CFC-113: IINFAR MODI I 2
NON-US OFCO USE 01 CFC-113: MNtAR MOOH. 2 IJIKIVN) I ROM U.S.
NON-OECD NON-AEROSOL USE OF CFC-113: LOG! T/POI'ULAT I ON MOOEL
USE
(AIL ESTIMATES IN Mil I IONS OF KILOGRAMS)
1985
1990
1995
2000
2(105
2010
2015
2020
2025
2030
2035
ANNUAL
GROW III
0,
5.
ti
it.
3
3.
2.
2.
?.
2,
2
2
?
2,
2
2
2
1
1
,0
9
.9
2
.5
2
9
,8
6
.8
.6
. 3
. 1
,0
.5
.2
. 0
.8
.6
U S
9.
587.
662.
738.
816.
933.
H)'l8.
1 161 .
12/2.
1381.
'>
U
.()
.9
.\
.9
o
,9
9
,2
!>
2
9
9
1
0
'*
!•>
.8
0.
0.
0,
0.
0.
0,
0.
0.
o.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
.0
, 0
.0
.0
, 0
.0
. 0
.0
.0
, 0
0
0
.0
,0
.0
.0
. 0
. 0
.0
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
,0
0
,0
. 0
.0
,0
.0
, 0
. 0
,0
.0
0
,0
0
,0
, 0
,0
.0
0
20/15
2050
2055
2060
2065
20/0
2075
ANNUAL AVI RAGE GROWTH F
-------
RESULTS FOR HIGH SCENARIO
-------
WORID JOTAL: CEC-I 1
OOMPONINIS INCIUDM) :
UNITED SIAIIS NON-AIROSOL USE 01 CIC-11: IINIAR MODI t 2
UNIIIU SI AlfS AtROSOL US! OK CIC-11: LINIAK MODH + A( ROSOL REST = 4.5
I EC NON-AEROSOL USE 01 CIC-11: LINEAR MODEL ?
EEC AIROSOL USE 01 CIC-11: IINfAR MODI L 2 + CIIKKINI Al ROSOI S RESIR.
OECD - US - EEC NON-AEROSOI USE 01 CIC-11: II Nl AK MODI I 2
01 CD - US - EEC AtROSOI USE 01 CIC-11: LINEAR MODH
NON-OECD NON-Al ROSOI USl 01 CfC-11: I OCI I/I'OI'Ul Al ION MODEL
NON-OICD AEROSOL USl 01 CEC-11: LOGI1/I'OI'UI Al ION MODI I
EMISSION EROM WORID BANK OF CEC-11: BESI ESMMATE - NO RESIRIC1IONS
(AIL ESI I MATES IN Ml I I IONS OE KILOGRAMS)
1985
1990
1995
2000
2OU5
2(110
2015
201'0
20?5
2050
20.15
2O40
2045
2050
2055
2060
2065
20/0
20/5
ANNUAL AVIRACE CROWIII RA1IS:
ANNUAI
CROWIII
0
5
5
5
6
5
5
5
'l
'1
3
2
2
0
/
l)
'1
0
8
6
3
8
2
'1
/
2
8
/
'1
3
2
1
U S
I|3D
5/4
/'>1
9//
1 115
1/51
2114
301 /
3849
4802
'jf'jl
6661
/4B9
8;'4 1
9009
il)92
1868
8
1
9
3
9
1
5
9
0
6
9
/
5
5
'4
21
21
20
20
14
14
3
3
2
2
2
2
2
2
2
2
2
2
2
. 1
. 1
.8
.8
.8
.8
.0
.0
.9
.9
.5
.5
.2
.2
. 1
. 1
.5
.5
.5
IBANK
818
123/
17/5
21)1)7
3 3 '40
«45')1
6066
798'4
H)3')2
13209
1657/
2036/
2')'l86
288')8
33U80
383/9
1)3490
48/77
54216
. 3
.8
. /
.8
.9
.3
.9
. 1
.3
.2
.2
.7
.8
. /
.4
. 1
.2
.6
.2
UBANK
91.6
616.9
1259. /
2035. 7
300O.6
4224.8
5/70.7
7702.9
100/6.0
1295/.5
16339.3
20142.5
242/3.5
28646.2
33288.5
38197.5
43320.0
48619.9
540/0.9
OBANK
726.6
620.9
515.9
412.1
340. 3
316.5
296.2
281 .2
266.3
251 . 7
237.9
225.2
213.3
202. 5
191.9
181.5
170.3
157.8
145.3
3. /53
3. /36
3.829
3. 745
0.0
-2.342
4. 770
7.346
-1.773
-------
WOKl0 I01AL: ClC-12
COHI'ONI NIS I NCI (11)1 I) :
UNHID SIAIIS NON-AIROSOI USI Ol CIC-12: I I Nl AK MODI I 2
UNI II I) SIAIIS Al ROSOI USI Ol CIC-12: I INI AH MODI I + Al ROSI RISI -- *4 *
IIC NON-AfROSOl USt Ol CIC-12: I INfAR MODH 2
IK: AIROSOL USI 01 CIC-12: MNIAR MODI I 2 + CIIRKINI Al ROSOI RCSIR.
OICD - US - IIC NON-AI HOSOI USI Ol CIC-12: IINIAR MODI I 2
01 CD - US - I I C Al ROSOI USI 01 CIC-12: MNIAR MODI I 2
NON-OtCD NON-AIROSOI USt Ol CIC-12: I OG I I /('Ol'lll Al I ON MODI I
NON-01 CD Af ROSOI USI Ol CIC-1?: I OGI I /I'OI'UI A I I ON MODI I
EMISSIONS H'>
2005
2010
2015
20,"'0
2025
2030
2035
20'IO
20'(5
2050
2055
21)60
2065
20/0
ANNUAI
GROW III
<).(>
6. 1
rj.fl
b.6
-3.1)
•). 1
'1.6
'(.0
3.'j
3.0
2.1)
2.0
. /
.5
.'j
.3
2
. 1
. 1
U S 1
6()H . 1
H2l|.r>
1 100. 1
HIM. \
1 ')1 '> . '1
2M/6.2
JI2H.9
3H'ji.5
'(61H.2
•>'l() 5 . 'l
6 1 'j'l . 2
6Hl>8. /
/'>!(>. 1
813:'. 9
8/9'».ll
9'l29.2
lOO'l ». f
1061(6. 1
1 121(1. !>
LMISSIONS
TOIAI
509.0
7'l / . 2
992 . 1
1323.6
175/.2
228 /. 3
2901. 1
3592. 1
'(326.8
50/'l.9
5811.0
6515.9
/180.U
7806.0
8'(2/.5
9()'l6.2
9659. /
10260.6
1085'(. /
{MISSIONS
1 ROM USI
'(3'(. 1
588. /
/91 . 5
1055.0
l'lO'(.l)
I82/.5
2318.9
2860.'!
i'(26.'(
3985.2
'15)3.5
5003. 8
51(56.5
58/6. f
628 /. 8
66/6. 3
/()'(9.2
/'(I 1 .9
//68.'l
1 MISSIONS
f ROM UlfANK
0.0
83. /
1 95 . 3
263. 1
3 '(6. 6
'(53.2
'j8«4. 1
/ 11 . /
900. '(
1089. /
1 29 / . '(
1512.0
1 /23.'(
1929. 5
21 !9.6
2 !69.8
26 10. '(
28'(8.6
3086.2
1 Ml SSIONS
IROM OBANK IBANK
/'1. 9 90'(.9
/'1. 9 13'(1.5
5.5 175/.8
5.5 2355.5
6 . 6 3086 . 2
6.6 3969.3
0.1 5018.0
0.
0.
0.
0.
o.
o.
0.
o.
0.
0.
0.
0.
625'(. 8
/65'(.9
9232.3
1092/.2
126'(5. 3
1'(3'(1 . 1
1599'(.'(
1 / /62 . 9
1 965 7 . /
215/5.9
23500.5
25'l32.3
UBANK
1 /'(.()
98'(.8
1636.8
2262. 3
3022.8
3939.0
500 / . 6
62'l'( . 8
/6'(5.'»
9223.2
10918.6
1263/. 1
1'(333.3
1 598 / . 1
1 //56.0
19651.3
215/0.0
23'(95.0
25'(2/.2
OBANK
731 .0
356.7
121 .0
93.3
63.'(
30.14
10.14
10.0
9.5
9. 1
8.6
8.2
7.7
7.3
6.9
6.'4
6.0
5.5
5. 1
ANNUAL AVI RAGL CROW III RAMS:
3.311
1. 29'l
3.'(58
3.25/
0.0
-7.221
3. 116
5.695
-5.366
-------
WORLD TOTAL: CfC-22
COMPONENTS I NCI UOEI) :
UNITED STATES TOTAL USE OE CFC-22: I INIAR MODII 2
NON-US OECD USE Ol CFC-22: LINEAR MODEE 2 01 HIVED fROM U.S. USE
NON-OECD NON-AEROSOI USE OF CFC-22: LOCIT/POPUI AT ION MODEL
(ALL ESTIMATES IN Ml I LIONS OE KILOGRAMS)
1985
1990
1995
2000
2005
2010
2015
2020
2025
2OJO
21135
20'tO
21 Hi 5
2050
2055
2060
2065
20/0
2(1/5
ANNUAL AVI RAGE GKOWIH KATIS:
ANNHAI
GROW III
0.
6
5
5
<4,
'1
'»
3
3
3
2
2
2
\
2
2
1
1
1
.0
.6
.7
. 1
.8
,»4
. 1
. 7
.3
.2
. 7
. 3
. 0
.8
.3
. 1
.9
. /
.6
U S
1 15
163
219
283
360
'l'49
552
666
/89
933
1076
1215
1 !5U
1'lHi
16 /'l
1H63
2051
2238
?'42'4
E
.2
.6
. 1
.»»
. <4
.8
.H
.6
.3
.(4
.2
.4
.9
.()
.9
.9
.U
.0
.6
EMISSIONS
TOIAl
51.
89.
136.
1 9'l .
266.
3'I3.
432.
530.
63/.
762.
892.
1026.
1161.
1296.
Hl61
1630.
18()'l.
198'4.
3169.
8
/
/
9
9
2
*4
,0
8
1
6
5
8
,8
. 1
3
, /
,8
, 1
EMISSIONS
EMISSIONS EMISSIONS
(ROM USf EROM UBANK E ROM OBANK
51.
73.
98.
127.
162.
202.
2'48.
30O.
355.
1420.
l|8<4.
5'4/.
607.
66 /.
/53.
838.
923.
1007.
1091 .
8
6
6
6
'1
i|
6
O
2
U
3
0
9
'4
/
8
1
1
1
0.
16.
38.
6/.
I0<4.
I'm.
183.
230.
282.
3'l2.
1408.
'4/9.
553.
629.
707.
/91.
881 .
9//.
10/8.
0
0
1
3
/
8
8
0
/
1
3
5
9
5
'4
6
6
/
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
TBANK
63.
»412.
808.
12'40.
1698.
2205.
27/8.
3'428.
»415/.
<4977.
58/3.
6810.
7/5/.
869'4 .
9/11.
108*12.
12052.
13308.
1*4581.
3
9
8
1
/
2
6
8
6
U
8
<4
'4
/
8
9
6
3
8
UBANK
63.
'412.
808.
12140,
1698.
2205.
2//8.
3*428.
»4157.
149/7,
5873.
6810.
7/57,
869<4,
9/11.
H)8'l2.
12052,
13308,
114581
3
,9
.8
. 1
7
2
6
8
.6
.14
8
.14
.'4
. /
.8
,9
.6
. 3
.8
OBANK
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
3.1456
14.236
3.<4'l'4
0.0
0.0
6.230
6.230
0.0
-------
WORID 10FAL: CFC-113
COMI'ONLNIS INCIUDH) :
UNIHO SIAItS 101AI USE OF CfC-113: I I NEAR MODI I 2
NON-US OtCD USE 01 CFC-113: UNEAR MOD El 2 PI l< I Vf 0 I ROM U.S. USE
NON-OECD NON-AEROSOL USE Of CFC-113: LOCI T/POI'ULAT ION MODEL
(AIL ESMMAFES IN Ml I I IONS OF KILOGRAMS)
% ANNUAL
GROW 1 II
1985
1 990
1995
2000
201)5
2010
2015
2(120
2025
2030
2(1 <5
2U'lO
20'l5
2050
2055
2060
2065
20/0
2075
0.
7.
5.
5.
i*.
b.
3.
3.
3.
3.
2.
2.
2.
1 .
2.
2.
2.
1.
1 .
0
0
8
1
I
3
9
6
2
2
7
i»
1
9
5
2
0
8
7
11 S
98.
1'|2.
192.
2') 8.
m.
391.
Ull.
575.
6/
.0
.0
.0
.()
. 0
. 0
.()
.0
.0
.0
.0
.0
.0
. o
. o
. 0
.0
0
0
0
u
0
0
0
0
0
o
0
0
0
0
0
0
0
0
0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
TBANK
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
o.o
0.0
0.0
0.0
0.0
0.0
0.0
UBANK
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
OBANK
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
ANNUAL AVERAGE GROWTH RAFES:
3.513
3.501
3.501
3 . 50 1
0.0
0.0
0.0
0.0
0.0
-------
RESULTS FOR NO LIMITS TO GROWTH SCENARIO
-------
WORLD 10IAL: CFC-11
COMI'ONI NIS I NCI 1)1)1 I) :
UN I III) SI AIIS NON-AI KOSOl t)SI 01 CIO 11: I INIAK MODI I 2
UNI I JO SI AIIS AIKOSOI USI Ol Cl C- 1 I : I I Nl AK MODI I + Al KOSOl RfSI l|. 5
LlC NON-AIROSOL USI Ol CIC-11: IINIAR MODI I 2
IIC Al IU)SOl USI Ol ClOU: IINIAK MODI I 2 + l.UKRINI Al KOSOl S KISIR.
Ol CD - US - IIC NON-AI KOSOl USI 01 CIC-11: I INIAK MODII 2
01 Cl) - US - IIC AIKOSOI USI 01 CIC-11: I INIAK MODII ?
NON-OI Cl) NON-AIKOSOI ItSI 01 CIC-11: I OG I I /I'OI'UI A I I ON MOON
NON-01 Cl) AIKOSOI USI Ol CIC-11: I OG I I / I'OI'UI A I I ON MODI I
I Ml SSI ON fKOM WOK I I) HANK 01 CIC-11: BIS I I S I I MA 11 - NO Kt SI K 1C I IONS
(All ISIIMAIIS IN Ml I I IONS OF KILOGRAMS)
% ANNUAI
GROW 1 II
I9H5
1990
1995
2OOO
2O05
2010
2015
21)20
2025
20 !O
2< t i 5
201)0
20')5
POM)
2055
2060
2065
20/0
20/5
0.
6.
6
6.
/
6
6
6.
5.
I).
'1.
3.
2.
2.
2.
2.
1 .
1 .
1 .
. 0
.6
.2
. 1
. S
. y
.5
.0
. 3
.9
.0
.2
6
;>
2
. 0
.8
6
.5
U S
518.
/2 !.
98 '1
1 12 /
1906 .
26/6.
3690.
'498 I .
6'>'lO
8'4/9.
1()'|85.
12'45'j.
I'l (29,
16096.
18110.
20065.
21992.
2391 /.
25858.
L
. /
, 5
. j
. 2
.8
.6
.9
y
6
. 3
8
, /
.8
5
. 7
. O
. 1
3
8
IMISSIONS
101AL
'426.
59 / .
82O.
Ill/,
1 596 .
225 / .
313/.
'12/0.
56'l / .
/ Cl6.
9128.
1O902.
12612.
1 ')2')l4 .
16036.
1 / /9'4 .
195'46.
2 1 309 .
23100.
3
.6
.2
.6
.6
.6
. 5
.8
.9
.6
5
.2
. /
/
/
.8
.2
5
'4
I Ml SSIONS
1 ROM USI 1
'105. 1
56 1 . 5
/63.6
1033. /
l')91 . .i
2 to/, i
2928. 1
(983. 1
5259.8
(.832.O
M'15/.B
IOOM /.8
1 1551 . 6
1/958. 8
1M508. O
15999.9
1 /'I62. /
1891 /. 9
20380.8
IMISSIONS IMISSIONS
ROM UBANK (ROM OBANK
O
15
(5
63
1OO
1 <|5
206
28'4
385
51 1
668
85 1
1O59
1283
1526
1/92
2081
2389
2/1 /
. O
.O
. 8
. 1
. •>
.6
.')
/
.2
.6
. ;>
.8
.0
. H
. /
.8
.()
. ;>
. I
21 .
21 .
20.
20.
'4.
l|
3'
3,
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
1
. 1
.8
.8
.8
.8
. 0
.0
.9
.9
, 5
•j
.2
2
1
1
. •>
5
. 5
IBANK
8')0,
I'lOl .
2 l')2 .
3096
'l'42'l .
6292
8//9
12015,
16113
21 300,
2/6<46.
3503i«.
'13305 .
5230'4
622'4l '.
/3209.
85O9'4.
9/815.
111309.
, 1
.2
.6
. 1
.2
/
. 0
/
.9
.3
.6
. 5
.'4
.6
.8
.2
. S
'4
. 1
UBANK
113
/80.
1626
268')
1)083
5 9/6
8'482
1 t/3'4
1 58'4 /
2IO'l8
2/'4()8,
3'l809
'43092,
52 1O2
6201)9,
7302/,
8'l92'l
9 /65 / .
1 1 1 163.
. 5
. 3
/
. 0
.9
. 3
.8
. ',
.6
. ',
. /
.2
. 1
. 0
.9
/
.2
, /
.9
OBANK
726
620
515
1)12
3 '40
316
296
281
266
251
237
225
213
202
191
181
1 70
157
1145
.6
.9
.9
. 1
.3
. 5
.2
.2
.3
.7
.9
.2
. 3
. ;>
.9
. *y
.3
.8
.3
ANNUAI AVI RAGl CKOWIII KAIIS:
'I.'l60 'I.'139
'4.536
'). '4Ml
0.0
-2.3'l2
/.952
-1.7/3
-------
WORLD 10TAL: CFC-12
COMPONtNIS INCIUDEO :
I1NIIU) SIAIES NON-AEKOSOI USF Of C1C-12: IINIAR MODI I 2
UNITIO SIATIS AEROSOl USE Of CEC-12: LINIAR MODEL + Al ROSI REST -- 4.5
I LC NON-AEROSOI USE Ol CIC-12: IINEAR MODEI ','
EEC AEROSOL USE Of CFC-12: I INtAR MODEL 2 + CURRENT Al ROSOl RESTR.
OECD - US - EEC NON-AtROSOL USE OE CFC-12: LINIAR MODEL 2
OECO - US - EFC AlROSOL USI 01 CIC-12: IINEAR MODEI 2
NON-OECD NON-AEROSOL USE 01 CIC-12: LOG I I/POPUI AT ION MODEI
NON-OfCD AEROSOI USE OF CfC-12: IOC I T/POI'Ul Al I ON MODI I
EMISSIONS FROM WORLD BANK OF CFC-12: BES1 ESIIMA1E - NO RLSIRICflONS
(AIL ESTIMATES IN Mil LIONS OF KIIOGRAMS)
ANNUAL
CROW III
0.0
6.9
6.'j
6.2
6.5
5
5
if
3
3
3
2
2
1
2
1
1
1
1
9
2
5
9
6
0
5
2
9
1
9
1
6
5
u s
/98
1 12H
1559
2119
29 43
3962
51/3
6'>4 )
801 /
969V
1 1359
1298U
14550
I6(l/ 1
1 /90/
19/26
21548
23 J84
25245
E
.2
. 3
.2
.2
.3
.6
.3
. /
. 7
. 1)
.'(
.8
.2
.9
.5
.8
. 1
.5
.14
EMISSIONS
TOTAL
618.2
996.8
1399.3
1924. •>
2670.2
3623
'1/59
6053
71)60
9U1 /
10616
12210
13766
1528'!
169'! lj
1867'j
20l»56
2?256/3
1 i /90
MI900
16012
1 /133
/
'4
6
6
>l
2
6
6
6
i
9
8
/
7
EMISSIONS EMISSIONS
ROM UISANK FROM OBANK IBANK
0.0 7'4.9 955.9
109.9 /U.9 1658.8
261.6 5.5 23'4l./
363.0 5.5 3242.6
'195.0 6.6 '4'437.5
6/6.
910.
1 183.
1501 .
18/3.
2308.
2/80.
3262.
3/'48.
'4?/2.
'IBS^.
5555.
62'4l.
69'45.
b
'>
1
/
3
0
/
/
8
D
o
/62.
'408 /'4.
'1625'!.
51830.
57591.
'4
2
0
1
0
2
<4
1
9
3
2
0
3
1
1985
1990
1995
2OOO
2005
2010
2015
2020
2025
2050
2035
20'lO
20'45
2050
2055
21)60
2065
20/0
20/5
ANNUAL AVI RACE GROWTH KA11S:
3.9(1 3.912
4.098
3. 8'4 /
0.0
-7.221
4.659
6.356
OBANK
731.0
356.7
121 .0
93.3
63.4
30.'4
10.4
10.0
9.5
9. 1
8.6
8.2
7.7
7.3
6.9
6.4
6.0
5.5
5. 1
-5.366
-------
WORLD TOTAL: CFC-22
COMF>ONtNTS INCLUDEO :
UNITED SIAtES TOIAI USE 01 CFC-22: UNFAR MODI I 2
NON-US OFCD USE OE CFC-22: LINEAR MODEI 2 DERIVED I ROM U.S. USE
NON-OFCO NON-AtROSOl USE OF CFC-22: LOGI F/POPUI AfION MODEL
(ALL ESTIMATES IN Ml I LIONS OE KILOGRAMS)
%
1985
1 990
1995
2000
2005
2010
2015
2020
2O25
2030
2O35
2O40
2045
2050
2O55
2O60
2O65
20/0
20/5
ANNUAF
GROW III
0.
7.
6.
5.
5.
5.
4.
4.
3.
3.
2.
2.
2.
3.
2.
2.
2.
2.
0
2
1
8
2
2
7
0
4
9
6
3
1
7
'l
2
0
U S
159.
231.
31 /.
418.
561.
/3 1 .
929.
1 154.
1400.
1 /2()
2()6o
239H.
2/39.
3082.
3622.
41 70.
4/29.
5300.
5885.
E
0
7
1
5
6
5
9
8
/
3
1
3
/
l)
/
j
9
6
EMISSIONS
TOTAL
/I.
126.
196.
284
405!
541.
705.
891 .
1103.
136/.
1655.
1962.
2284.
2618.
3055.
3522.
4020.
4550.
5108.
6
6
1
9
9
/
5
8
0
1
2
2
7
5
2
5
5
7
4
EMISSIONS
EMISSIONS EMISSIONS
I ROM USE FROM UBANK FROM OBANK
/I .
104.
142.
188.
252.
329.
418.
519.
630.
/ /6
92/.
10/9.
1232.
I38/.
1630.
18/6.
2128.
2385.
2648.
6
3
I
3
I
2
4
I
3
8
2
l|
f
2
1
8
4
i|
5
0.0
22.4
54.0
96.6
153.2
212.5
28/.0
3/2.2
4/2. /
590 . 3
728.0
882. /
1052.0
1231. 5
1425. 1
1645. /
1892. 1
2165. 3
2459.9
0.
0.
0.
0.
0.
0.
0.
o .
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0
0
0
0
0
0
()
0
0
0
0
0
0
0
0
0
0
0
0
TBANK
8 / .
578.
1 150.
1/93.
2528.
3408.
4461.
5/00.
7121 .
8800.
10740.
12866.
15108.
1/411.
20050.
23137.
26569.
30246.
34080.
5
5
8
3
6
8
9
5
4
1
4
4
0
1
3
3
3
5
4
UBANK
8/.5
578.5
1 150.8
1/93.3
2528.6
3408.8
4461 .9
5/00.5
7121.4
8800. 1
10/40.4
12866.4
15108.0
1/411.1
20050. 3
2313/. 3
26569.3
30246.5
34080.4
ANNUAL AVI KAGF GROWTH ISAFIS:
'I. 106
'4.094
14.856
4.094
0.0
0.0
6.852
6.852
OBANK
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
-------
WORI i> loiAL: CM:-in
COHI'ONI Nl S I NCI 1)01 I) :
UN I I II) SI AMS IOIAI USI Ol CIC-1U: IINIAK MODI I 2
NON-US 01 CO USE 01 CK:-113: I I Nl AH MOOI I 2 1)1 l< I VI D I KOM U S USI
NON-01 CD NON-AEKOSOI USE 01 CEC-113: LOG I T/POI'ULAT I ON MOOI L
(AIL ESI I MATES IN Ml I I IONS Of KILOGRAMS)
%
1985
1990
1995
?0l)0
2005
POIO
2015
2020
2025
20 so
20 1'>
?()'io
20'45
?(!')()
2055
2OM»
2065
20/0
20/5
ANNUAL
CROW III
0.0
1.6
6.3
5.5
5.8
5.2
'I. /
14.2
3. /
1.0
3. '4
3.0
2.6
2.14
3.3
2.9
2.6
2.3
2. 1
U S E
1211.5
18/. 1
259 . 8
3'4'4.8
'46'4. 3
6()'l . 5
/66 . 3
9'48.5
1 1 'I / . 2
1 '11 H . /
U>99.()
198'l.8
?:J /'».()
2569.8
3l)'l9.5
3538.6
'I03H.5
'4 5 5(1. '4
50/i|.9
EMISSIONS
TOTAL
1 2'4 . 5
187. 1
259.8
3'|l4.8
14614.3
60»4.5
/66.3
9'48.5
1 1 '4 / . 2
1U18. 7
1699.0
198'4.8
2275.0
2569.8
30119.5
3538.6
'4038.5
'4550.'4
5074.9
LMISSIONS EMISSIONS EMISSIONS
II«)M IJSL fROM UBANK FROM OBANK
12'4.
18/.
259.
3'l'4.
«46'4.
60'4.
766.
9'48.
11 'I/.
1 '4 1 8 .
1 699 .
198'4.
22/5.
2569 .
30'49.
.5538.
'1038.
'1550.
50/4.
5
1
8
8
3
5
3
5
2
/
0
8
0
8
5
6
5
'4
9
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
O.
0.
0.
0.
0.
0.
0.
0
I)
O
0
0
I)
O
0
O
0
O
O
0
I)
O
O
0
0
0
0.0
0.0
0.0
0.0
o.o
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
FBANK
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
o.o
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
UBANK
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
o.o
0.0
0.0
0. 0
0.0
0.0
0.0
0.0
o.o
0.0
0.0
OBANK
0.0
0. 0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0. 0
0.0
0.0
0. 0
0.0
0.0
0.0
0.0
ANNUAL AVI KAGE GKOWTII KATIS:
4.219 '4.206
14.206
14.206
0.0
0.0
0.0
0.0
-------
ANNEX E
POPULATION AND GNP PER CAPITA PROJECTIONS
Projections of population and GNP per capita in the U.S., non-U.S. OECD,
and the non-OECD countries were developed to provide a basis for projecting
scenarios of potential future CFC use. Five projections were developed:
(1) "Limits to Growth," a likely lower bound on future GNP
per capita and population growth;
(2) Low;
(3) Medium;
(4) High; and
(5) "No Limits to Growth," a likely upper bound on future
GNP per capita and population growth.
Exhibit E-l displays the population projections. First, global population
projections were adopted from a review of published estimates; these are shown
at the top of Exhibit E-l. Edmunds' (1984) low scenario was the lowest
published projection, only 7.1 billion people in 2075.1 For this population
to be achieved, all portions of the world must achieve zero population growth
by the middle of the next century.
The highest published estimate is from the United Nations (1981), over
13.6 billion people in 2075.2 The middle three projections were developed
from Lovins (1981),3 the World Bank (1985)," and Edmunds' (1984)5 middle
scenario. To divide these global estimates into three regions, the expected
share of world population in each region over time was estimated from Edmunds'
three scenarios.6 These shares (shown in Exhibit E-l) are
1 J.A. Edmunds and J. Reilly, et al. An Analysis of Possible Future
Atmospheric Retention of Fossil Fuel CO . Prepared for United States
Department of Energy, Washington, D.C. September, 1984.
2 Long-Range Global Population Projections 1984. World Bank,
Washington, D.C., 1985.
3 Amory Lovins et al. Least Cost Energy: Solving the C00 Problem,
Brick House Publishing Company, Andover, MA. 1981.
u My T. Vu. World Population Projections 1984. World Bank,
Washington, B.C., 1985.
5 Edmunds, op. cit.
6 Ibid.
-------
E-2
EXHIBIT E-1
GLOBAL POPULATION PROJECTIONS
(MILLIONS)
YEAR
1985
2000
2025
2050
2075
LIMITS
TO GROWTH
4536
5377
6505
7324
7131
LOW MEDIUM
4745 4745
5901 5901
7384 7384
7664 8223
7944 8491
REGIONAL POPULATION SHARES
HIGH
4835
6147
8160
9496
9960
NO LIMITS
TO GROWTH
5000
6500
9500
12100
13600
YEAR
US SHARE
OECD-US SHARE
NON-OECD SHARE
1985 0.048
2000 0.043
2025 0.038
2050 0.036
2075 0.035
0.120
0.107
0.094
0.088
0.087
0.832
0.850
0.868
0.876
0.879
REGIONAL POPULATION PROJECTIONS
YEAR REGION
1985 US POP
OECD-US POP
NON-OECD POP
TOTAL
2000 US POP
OECD-US POP
NON-OECD POP
TOTAL
2025 US POP
OECD-US POP
NON-OECD POP
TOTAL
2050 US POP
OECD-US POP
NON-OECD POP
TOTAL
2075 US POP
OECD-US POP
NON-OECD POP
TOTAL
LIMITS
TO GROWTH
219
545
3772
4536
232
576
4569
5377
249
612
5644
6505
262
644
6418
7324
247
617
6267
7131
(MILLIONS)
LOW
229
570
3946
4745
255
632
5014
5901
283
695
6407
7384
274
674
6716
7664
275
688
6981
7944
MEDIUM
229
570
3946
4745
255
632
5014
5901
283
695
6407
7384
294
723
7206
8223
294
735
7462
8491
HIGH
234
581
4020
4835
266
658
5223
6147
313
768
7080
8160
339
835
8322
9496
345
862
8753
9960
NO LIMITS
TO GROWTH
242
601
4158
5000
281
696
5523
6500
364
894
8243
9500
432
1064
10604
12100
471
1177
11951
13600
-------
E-3
multiplied by the global values to produce population projections by region
over time. As indicated by the increasing share of world population in
non-OECD countries, this region's population is projected to grow faster than
the other regions'.
To project GNP per capita we reviewed projections and historical data on
the rates of growth of GNP and GNP per capita. Historically, the rates of
growth of these quantities have varied significantly across regions and
throughout time. The largest reported sustained growth of GNP per capita is
for Japan, 1874 to 1967, 2.8 percent per year.7 It is unlikely that global
GNP per capita will grow this rapidly over the next 90 years. The lowest
projected rates of growth are in Lovins (1981),8 with growth declining from
one percent per year through 2000, to under 0.3 percent per year by 2075. For
purposes of developing projections, we assumed that all our scenarios would
fall within these wide boundaries. To compute the five scenarios, we assumed
the following growth rates in GNP per capita:
Limits to Growth
Low
Medium
High
No Limits to Growth
1985-2000
1.0%
1.5%
2.0%
2.5%
2000-2025
0.5%
1.2%
1.7%
2.0%
2.5%
2025-2050
0.5%
1.0%
1.7%
2.0%
2.5%
2050-2075
0.3%
1.0%
1.7%
2.0%
2.5%
These values reflect the full range of likely global GNP per capita over the
next 90 years.
To develop estimates for each of the three regions of interest, we
performed the following 5 steps:
(1) compute global GNP per capita through 2075 for each of
the five projections (see top of Exhibit E-2);
(2) compute global GNP by multiplying the global GNP per
capita by the number of people estimated in Exhibit E-l
(see middle of Exhibit E-2);
Simon Kuznets. Economic Growth of Nations, The Beltnap Press of
Harvard University Press, Cambridge, MA., 1972.
8 Lovins, op. cit.
-------
E-4
EXHIBIT E-2
PROJECTIONS OF GLOBAL
YEAR
1985
2000
2025
2050
2075
LIMITS
TO GROWTH
1900
2206
2499
2831
3051
GNP PER
(1975
LOW
1900
2375
3201
4105
5264
GLOBAL
COMPUTED USING GNP PER
YEAR
1985
2000
2025
2050
2075
YEAR
1985
2000
2025
2050
2075
LIMITS
TO GROWTH
8618
11861
16254
20731
21755
US SHARE
0.227
0.215
0.179
0.159
0.162
(BILLIONS OF
LOW
9016
14017
23635
31459
41818
REGIONAL
OECD-US
0.
0.
0.
0.
0.
CAPITA
us $)
MEDIUM
1900
2447
3729
5683
8662
GNP
CAPITA AND
1975 US S)
MEDIUM
9016
14438
27535
46735
73553
GNP SHARES
SHARE
391
387
374
365
369
NO
HIGH TO
1900
2557
4195
686.3
11292
POPULATION
NO
HIGH TO
9187
15719
34233
65359 1
LIMITS
GROWTH
1900
2752
5102
9458
17535
LIMITS
GROWTH
9500
17886
48465
14443
112468 238474
NON-OECD
0.
0.
0.
0.
0.
SHARE
381
399
447
475
469
-------
EXHIBIT E-2 (continued)
REGIONAL GNP PROJECTIONS
YEAR
1985
2000
2025
2050
2075
REGION
US GNP
OECD-US GNP
NON-OECD GNP
TOTAL
US GNP
OECD-US GNP
NON-OECD GNP
TOTAL
US GNP
OECD-US GNP
NON-OECD GNP
TOTAL
US GNP
OECD-US GNP
NON-OECD GNP
TOTAL
US GNP
OECD-US GNP
NON-OECD GNP
TOTAL
(BILLIONS
LIMITS
TO GROWTH
1959
3372
3287
8618
2547
4584
4730
11861
2909
6074
7272
16254
3303
7573
9855
20731
3520
8023
10211
21755
OF 1975
LOW
2049
3528
3439
9016
3010
5418
5590
14017
4229
8832
10573
23635
5012
11492
14955
31459
6766
15423
19629
41818
U.S. S)
MEDIUM
2049
3528
3439
9016
3100
5580
5757
14438
4927
10289
12318
27535
7446
17072
O "> "> 1 "*
46735
11901
O ~* 1 1 "•
_ i 1 .. /
34525
73553
HIGH
2088
3595
3504
9187
3375
6075
6268
15719
6126
12792
15315
34233
10413
23575
31071
65359
18198
-+1-+79
52791
112-468
NO LIMITS
TO GROWTH
2159
3717
3624
95GO
3841
6913
"™ 1 "^ *X
17886
8673
18111
21682
48465
18234
-130-
5-*-C5
11-+-43
355S"
S'95C
111937
23347-
-------
E-6
(3) estimate regional shares of global GN'P from Edmunds'
(1984) estimates9 (see middle of Exhibit E-2);
(4) multiply the regional GNP shares by the global GNP
estimates to compute regional GNP estimates (see end of
Exhibit E-2); and
(5) divide the regional GNP projections by the regional
population projections (Exhibit E-l) to compute
regional GNP per capita projections (see Exhibit E-3).
The values shown in Exhibit E-3 show GNP per capita in the U.S. growing at an
average rate between 0.5 percent and'2.5 percent per year. The rates in the
non-OECD countries are larger: 0.7 percent to 2.7 percent per year.
Edmunds , OJD •
-------
E-7
EXHIBIT E-3
REGIONAL GNP PER CAPITA PROJECTIONS
YEAR
1985
2000
2025
2050
2075
REGION
US GNP
OECD-US GNP
NON-OECD GNP
TOTAL
US GNP
OECD-US GNP
NON-OECD GNP
TOTAL
US GNP
OECD-US GNP
NON-OECD GNP
TOTAL
US GNP
OECD-US GNP
NON-OECD GNP
TOTAL
US GNP
OECD-US GNP
NON-OECD GNP
TOTAL
(1975
LIMITS
TO GROWTH
8932
6189
872
1900
10965
7964
1035
2206
11674
9927
1288
2499
12625
11758
1536
2831
14243
12998
1629
3051
U.S. $)
LOW
8932
6189
872
1900
11808
8577
1115
2375
14953
12716
1650
3201
18308
17051
2227
4105
24576
22429
2812
5264
MEDIUM
8932
6189
872
1900
12162
8834
1148
2447
17421
14814
1923
3729
25350
23609
3083
5683
40441
36908
4627
8662
HIGH
8932
6189
872
1900
12711
9233
1200
2557
19599
16667
2163
4195
30699
28591
3734
6883
52718
48112
6031
11292
NO LIMITS
TO GROWTH
8932
6189
872
1900
13678
9935
1291
2752
23834
20268
2630
5102
42186
39289
5131
9458
81863
74711
9366
17535
-------
R-3386-EPA
Product Uses and
Market Trends for
Potential Ozone-
Depleting Substances,
1985-2000
James K. Hammitt, Kathleen A. Wolf,
Frank Camm, William E. Mooz,
Timothy H. Quinn, Anil Bamezai
May 1986
Prepared for the
U.S. Environmental Protection Agency
-------
PREFACE
This report is one of a series of papers written at The Rand Cor-
poration on policy issues associated with chemicals that could poten-
tially deplete ozone in the stratosphere ("potential ozone depleters").
Stratospheric ozone is important because the ozone layer helps shield
the earth from harmful ultraviolet radiation. Increases in ultraviolet
radiation may threaten human health, speed deterioration of certain
materials, reduce crop yields, and have a wide range of potentially
important ecological effects. Atmospheric models developed and tested
over the last decade suggest that global human emissions of potential
ozone depleters may lead to chemical reactions that reduce strato-
spheric ozone, thereby increasing ultraviolet radiation with its concom-
itant effects. Substantial scientific uncertainty persists about whether
human emissions of these chemicals actually threaten the stratospheric
ozone layer and, if they do, whether lower ozone levels actually
threaten human health and other activities at the earth's surface that
concern policymakers. Policymakers must act in the face of this uncer-
tainty, however, and Rand's work is designed to help them act with the
best information available.
To that end, The Rand Corporation is developing a series of reports
addressed to analysts and policymakers responsible for policy decisions
on emissions of potential ozone depleters in the United States and else-
where. These documents report the results of research that includes
extensive literature reviews, interviews with knowledgeable officials
associated with the production and use of potential ozone depleters,
and formal chemical, cost, economic, and statistical analyses. The
series should also interest the much broader audience of analysts and
decisionmakers whose organizations would feel the effects of govern-
ment policies with respect to emissions of such chemicals.
Published papers in the series include the following:
• A. R. Palmer, W. E. Mooz, T. H. Quinn, and K. A. Wolf,
Economic Implications of Regulating Chlorofluorocarbon Emis-
sions from Nonaerosol Applications, R-2524-EPA, June 1980.
• A. R. Palmer, W. E. Mooz, T. H. Quinn, and K. A. Wolf,
Economic Implications of Regulating Nonaerosol Chlorofluorocar-
bon Emissions: An Executive Briefing, R-2575-EPA, July 1980.
• K. A. Wolf, Regulating Chlorofluorocarbon Emissions: Effects on
Chemical Production, N-1483-EPA, August 1980.
111
-------
IV
• A. R. Palmer and T. H. Quinn, Economic Impact Assessment of
a Chlorofluorocarbon Production Cap, N-1656-EPA, February
1981.
» A. R. Palmer and T. H. Quinn, Allocating Chlorofluorocarbon
Permits: Who Gains, Who Loses, and What Is the Cost?
R-2806-EPA, July 1981.
• W. E. Mooz, S. H. Dole, D. L. Jaquette, W. H. Krase, P. F.
Morrison, S. L. Salem, R. G. Salter, and K. A. Wolf, Technical
Options for Reducing Chlorofluorocarbon Emissions,
R-2879-EPA, March 1982.
• E. M. Sloss and T. P. Rose, Possible Health Effects of Increased
Exposure to Ultraviolet Radiation, N-2330-EPA, July 1985.
• T. H. Quinn, K. A. Wolf, W. E. Mooz, J. K. Hammitt, T. W.
Chesnutt, and S. Sarma, Projected Use, Emissions, and Banks
of Potential Ozone-Depleting Substances, N-2282-EPA, January
1986.
• F. Camm and J. K. Hammitt, An Analytic Method for Con-
structing Scenarios from a Subjective Joint Probability Distribu-
tion, N-2442-EPA, May 1986.
• F. Camm, T. H. Quinn, A. Bamezai, J. K. Hammitt, M.
Meltzer, W. E. Mooz, and K. A. Wolf, Social Cost of Technical
Control Options to Reduce the Use of Potential Ozone Depleters
in the United States: An Update, N-2440-EPA, May 1986.
• W. E. Mooz, K. A. Wolf, and F. Camm, Potential Constraints on
Cumulative Global Production of Chhrofluorocarbons,
R-3400-EPA, May 1986.
This report was produced under Cooperative Agreement No.
CR811991-02-0 with the U.S. Environmental Protection Agency.
-------
SUMMARY
Global human emissions of chlorofluorocarbons (CFCs) and several
related chemicals may reduce the concentration of stratospheric ozone
and, in so doing, may induce significant negative effects on human
health and a variety of economically important activities. Emissions of
these "potential ozone depleters" depend fundamentally on their pat-
terns of production over time. For the United States and the world as
a whole, this report describes current patterns of use and projects
future production levels for the seven most important potential ozone
depleters—CFC-11, CFC-12, CFC-113, carbon tetrachloride, methyl
chloroform, Halon 1211, and Halon 1301—over the period 1985 to
2000. The projections are based on detailed analysis of the major
applications of these chemicals. The analysis is particularly detailed
for CFC-11 and CFC-12, which are believed to be the most important
potential ozone depleters.
Table S.I summarizes the findings on estimated current and pro-
jected use for the world as a whole. Each projection includes a
midrange estimate and a range of uncertainty around this estimate for
2000 that reflects three sources of uncertainty:
Table S.I
ESTIMATED CURRENT AND PROJECTED WORLD USE OF
POTENTIAL OZONE-DEPLETING SUBSTANCES
Use
(thousands of metric tons)
Average Annual
Projected in 2000
Chemical
CFC-11
CFC-12
CFC-113
Methyl chloroform
Carbon tetrachloride
Halon 1301
Halon 1211
1985
341.5
443.7
163.2
544.6
1029.0
10.8
10.8
Midrange
560
620
420
840
1550
20
20
Uncertainty (%)
-25 to +32
-26 to +34
-31 to +44
-26 to +34
-25 to +33
-38 to +59
-39 to +60
Growth
Lower
1.4
0.3
3.9
0.9
0.8
1.1
0.9
Rate (%)
Upper
5.2
4.3
9.2
5.0
4.8
7.6
7.6
-------
VI
• The rate of general economic growth,
• The rate of growth in demand for products using each chemical
relative to the rate of general economic growth, and
• Economic, technical, and regulatory effects on demand for each
chemical relative to the demand for the products in which it is
used.
The projections assume no change in the regulations of these sub-
stances relating to potential ozone depletion, although they do consider
other regulations such as the U.S. ban on land disposal of waste chlori-
nated solvents.
CFC-11 and CFC-12 are used in a variety of applications. Both are
widely used as aerosol propellants. CFC-11 is used extensively as a
blowing agent in manufacturing plastic foams and CFC-12 is widely
used as a refrigerant. Carbon tetrachloride's principal use is as a
chemical intermediate in the production of CFC-11 and CFC-12,
although it is also used as a solvent and grain fumigant. CFC-113 and
methyl chloroform are both solvents. CFC-113 is widely used in the
electronic industry, whereas methyl chloroform is a general purpose
solvent used in everything from electronics to shipbuilding. Halon
1301 and 1211 are newly marketed fire extinguishants used to protect
valuable equipment.
In general, production of these chemicals will likely grow in step
with the global economy, but average annual growth rates for individ-
ual chemicals could range from near 0 to 8 and 9 percent over the next
15 years. CFC-113 and the Halons are likely to grow the fastest.
Atmospheric modelers could use subjective probability distributions
for chemical production, such as those reflected in the limits on the
production ranges shown in Table S.I, to construct emission scenarios.
In doing this, they should keep in mind that the joint probability of
observing all low or all high growth rates is far smaller than the proba-
bility of observing a low or high growth rate for any chemical individu-
ally. Moreover, the growth rates for each chemical are probably corre-
lated because of their relationship to general economic growth. To
construct emission scenarios that account for the correlation among
individual chemical growth rates, we propose a method and a resulting
set of production scenarios in Camm and Hammitt (1986). By taking
account of the time pattern of emissions from various applications
these production scenarios can be translated to emission scenarios
similar to those reported in Quinn et al. (1986).
In assessing the importance of each chemical to the possibility of
significant depletion of stratospheric ozone (or other concerns, like
climatic change), it is not appropriate to simply compare production
-------
vn
levels. Both the fraction of production that is ultimately emitted and
the time pattern of emissions vary among chemicals. Moreover, the
estimated effect of an equal quantity of each chemical on the ozone
differs dramatically, perhaps by a factor of one hundred or more for the
seven potential ozone depleters we consider. Thus, even though carbon
tetrachloride and methyl chloroform are produced in the largest quanti-
ties, CFC-11 and CFC-12 are probably the most important of the
chemicals considered here. Consequently, the report examines their
use in the greatest detail.
The development of the projections reported here involves the
development of a transparent set of accounts that can be used to calcu-
late the joint implications of a large number of necessarily subjective
judgments about the future. Wherever possible, we have sought the
best expert judgment available as a basis for these projections. In the
end, of course, the reader may disagree with judgments reported here.
To the full extent possible, we have attempted to make it as simple as
possible to trace the implications of changes in the assumptions that
underlie the current results.
The accounts developed differ from one chemical to another to suit
differences in the quality of data available. The production of CFC-11
and CFC-12 and U.S. chemical production generally receive the most
attention because data on other chemicals and regions are in general
less complete. Other world production outside the communist coun-
tries receives the next most detailed treatment, although gaps in the
available data often require that we rely on U.S. data to infer produc-
tion levels elsewhere. Production in the communist countries is the
most difficult to analyze. To estimate global totals we estimate com-
munist country use based on the best available data, but these esti-
mates are the least certain of all.
In general, the accounts work from the bottom up, building from
information about chemical use in particular products and regions to
global production. But this method does not always yield global totals
that are consistent with "top-down" information from other sources,
like the Chemical Manufacturers Association. We review our differ-
ences with top-down sources and suggest a way to reflect them in the
projections developed here.
Rand's work on current and likely future production of these chemi-
cals is continuing, particularly with regard to U.S. production.
Although the information presented here represents the best informa-
tion we have today, we expect our understanding of current and
future trends to improve as the work proceeds.
-------
ACKNOWLEDGMENTS
We are indebted to many individuals and organizations for assis-
tance in writing this report. Much of the information contained in the
report was obtained through intensive interaction with industry
representatives beginning in the 1970s. Over this period a large
number of companies and trade organizations cooperated with Rand by
providing basic data, current estimates, and projections. Many of these
organizations are cited within the report: Others wish to remain
anonymous. We thank them all.
Rand colleagues Arthur Alexander and Richard Salter painstakingly
reviewed earlier drafts, David Rubenson and Toshiya Hayashi collected
information on refrigeration applications, and Jan Acton helped to
hasten review and production of the report under stringent deadlines.
John Hoffman and Stephen Seidel of the U.S. Environmental Protec-
tion Agency helped to frame the analysis and contributed numerous
specific suggestions. In addition, we received valuable comments from
the Alliance for Responsible CFC Policy, John Wells, and other partic-
ipants at the EPA-sponsored workshop "Protecting the Ozone Layer:
Workshop on Demand and Control Technologies." Mary Vaiana
assisted in preparing the presentation of the material to that workshop,
Patricia Bedrosian ably edited the final draft, and Janet D'Amore, Chai
Fosterling, Nancy Lees, and Alyce Shigg assisted in preparation of the
manuscript.
Especially in a research area as contentious as this, none of the indi-
viduals and organizations that assisted us can be presumed to agree
with all our conclusions. As always, the authors alone are responsible
for any judgments or remaining errors of fact.
IX
-------
CONTENTS
PREFACE iii
SUMMARY v
ACKNOWLEDGMENTS ix
FIGURES xiii
TABLES xv
Section
I. INTRODUCTION 1
Current Applications 4
Summary of Projected Use 7
Organization of the Report 11
II. METHODOLOGY 12
Characterization of the Uncertainty About Projected
Chemical Use 12
Projected Growth in GNP 14
III. AEROSOL PROPELLANTS 17
IV. RIGID FOAM 22
Rigid Urethane Foam 23
Nonurethane Foam 26
V. FLEXIBLE FOAM 31
Slabstock Foam 31
Molded Foam 34
VI. REFRIGERATION AND AIR CONDITIONING
SYSTEMS 39
Mobile Air Conditioning 39
Retail Store Refrigeration 42
Home Refrigerators and Freezers 44
Chillers 47
VII. MISCELLANEOUS USES OF CFC-11 AND CFC-12 54
Sterilants 54
Liquid Food Freezing 54
Other Miscellaneous Applications 55
XI
-------
Xll
VIII. SOLVENTS 57
CFC-113 58
Methyl Chloroform 61
Carbon Tetrachloride 63
IX. FIRE EXTINGUISHANTS 68
Historical U.S. Use 69
Current and Future Reporting Country Use 70
X. USE OF POTENTIAL OZONE DEPLETERS IN THE
COMMUNIST COUNTRIES 74
CFC-11 and CFC-12 Use 74
Solvent Use 76
XL CONCLUSIONS 79
United States and Global Projections 79
Directions for Future Work 84
Appendix
A. ESTIMATES OF CURRENT CONSUMPTION OF
CFC-11, CFC-12, METHYL CHLOROFORM, AND
CARBON TETRACHLORIDE 87
B. COMPARISON OF ESTIMATED TOTAL CFC-11
AND CFC-12 USE WITH OTHER SOURCES 93
C. DERIVATION OF SUBJECTIVE CREDIBILITY
INTERVALS 101
REFERENCES 105
-------
FIGURES
1.1. Estimated CMA Reporting Country and U.S. Use of
CFC-11, by Product 5
1.2. Estimated CMA Reporting Country and U.S. Use of
CFC-12, by Product 6
3.1. Estimated Historical and Projected U.S. Use of CFC-11
and CFC-12 in Aerosols 19
4.1. Estimated Historical and Projected U.S. Use of CFC-11
and CFC-12 in Rigid Foam 29
4.2. Estimated Historical and Projected Reporting Country
Use of CFC-11 and CFC-12 in Rigid Foam 30
5.1. Estimated Historical and Projected U.S. Use of CFC-11
in Flexible Foam 37
5.2. Estimated Historical and Projected Reporting Country
Use of CFC-11 in Flexible Foam 38
6.1. Estimated Historical and Projected U.S. Use of CFC-11
and CFC-12 in Refrigeration and Air Conditioning 52
6.2. Estimated Current and Projected Reporting Country
Use of CFC-11 and CFC-12 in Refrigeration and Air
Conditioning 53
8.1. Estimated Historical and Projected U.S. Use of
CFC-113, Methyl Chloroform, and Carbon
Tetrachloride 66
8.2. Estimated Current and Projected Reporting Country
Use of CFC-113, Methyl Chloroform, and Carbon
Tetrachloride 67
9.1. Estimated Historical and Projected U.S. and Reporting
Country Use of Halon 1301 and Halon 1211 73
10.1. Estimated Historical and Projected Use of CFC-11 and
CFC-12 in the Communist Countries 77
11.1. Estimated Historical and Projected Use of CFC-11 and
CFC-12 in the United States 81
11.2. Estimated Historical and Projected Use of CFC-11 and
CFC-12 in the Reporting Countries 82
Xlll
-------
TABLES
S.I. Estimated Current and Projected World Use of
Potential Ozone-Depleting Substances v
1.1. Estimated 1985 World Use of Potential Ozone-Depleting
Substances 2
1.2. Estimated Current and Projected World Use of Potential
Ozone Depleters 8
1.3. Estimated Current and Projected World Use of CFC-11 . . 9
1.4. Estimated Current and Projected World Use of CFC-12 . . 10
2.1. Projected Base GNP Growth Rates 15
3.1. Estimated Historical Use of CFC-11 and CFC-12
in Aerosols 18
3.2. Estimated Current and Projected Use of CFC-11 and
CFC-12 in Aerosols 20
4.1. Estimated Historical U.S. Production of Rigid Urethane
Foam 23
4.2. Estimated Historical U.S. Use of CFC-11 in Rigid
Urethane Foam 24
4.3. Estimated Historical Reporting Country Use of CFC-11
in Rigid Urethane Foam 24
4.4. Estimated Current and Projected U.S. Use of CFC-11
in Rigid Urethane Foam 25
4.5. Estimated Current and Projected Reporting Country
Use of CFC-11 in Rigid Urethane Foam 26
4.6. Estimated Historical U.S. Production of Nonurethane
Foam and Use of CFC-12 27
4.7. Estimated Current and Projected U.S. Use of CFC-12
in Nonurethane Foam 28
4.8. Estimated Current and Projected Reporting Country
Use of CFC-12 in Nonurethane Foam 28
5.1. Estimated Historical U.S. Production of Slabstock
Foam and Use of CFC-12 32
5.2. Estimated Current and Projected U.S. Use of CFC-11
in Slabstock Foam 33
5.3. Estimated Historical Reporting Country Use of CFC-11
in Slabstock Foam 34
5.4. Estimated Current and Projected Reporting Country
Use of CFC-11 in Slabstock Foam 34
5.5. Estimated Historical U.S. Production of Automotive
Vehicles, Molded Foam and Use of CFC-11 35
XV
-------
XVI
5.6. Estimated Current and Projected Reporting Country
Automotive Vehicle Production and Use of CFC-11
in Molded Foam 36
6.1. Estimated Historical and Projected U.S. Vehicle Sales
and Use of CFC-12 in Mobile Air Conditioning 41
6.2. Estimated Current and Projected Reporting Country Use
of CFC-12 in Mobile Air Conditioning 41
6.3. Estimated Historical and Projected U.S. Use of CFC-12
in Retail Food Refrigeration 43
6.4. Estimated Current and Projected Reporting Country
Use of CFC-12 in Retail Food Refrigeration 44
6.5. Estimated Historical and Projected U.S. Home Refrigerator
and Freezer Sales and Use of CFC-12 45
6.6. Estimated Historical World Production of Home
Refrigerators 46
6.7. Estimated Current and Projected Reporting Country
Use of CFC-12 in Home Refrigerators and Freezers 47
6.8. Estimated Historical and Projected U.S. Centrifugal
Chiller Installations and Use of CFC-11 and CFC-12 .... 48
6.9. Estimated Current and Projected Reporting Country Use of
CFC-11 and CFC-12 in Centrifugal Chillers 49
6.10. Estimated Historical and Projected U.S.
Reciprocating Chiller Installations and Use of CFC-12 ... 50
6.11. Estimated Current and Projected World Use of CFC-12
in Reciprocating Chillers 51
7.1. Estimated Current and Projected Use of CFC-12 in
Liquid Food Freezing 55
7.2. Estimated Current and Projected Use of CFC-11 and
CFC-12 in Other Miscellaneous Applications 56
8.1. Estimated Historical and Projected U.S.
Use of CFC-113 59
8.2. Estimated Current and Projected Reporting Country Use
of CFC-113 60
8.3. Estimated Historical and Projected U.S. Use of Methyl
Chloroform 62
8.4. Estimated Current and Projected Reporting Country
Use of Methyl Chloroform 62
8.5. Estimated Historical U.S. Use of Carbon
Tetrachloride 64
8.6. Estimated Current and Projected Reporting Country
Use of Carbon Tetrachloride 64
9.1. Estimated Historical U.S. Use of Halon 1301 69
-------
XV11
9.2. Estimated Current and Projected Reporting Country
Use of Halon 1301 71
9.3. Estimated Current and Projected Reporting Country
Use of Halon 1211 71
10.1. Estimated Historical Soviet Union and Projected
Communist Country Use of CFC-11 and CFC-12 75
10.2. Estimated Current and Projected Communist Country
Use of CFC-113, Methyl Chloroform, and Carbon
Tetrachloride 78
11.1. Estimated Current and Projected U.S. Use of Potential
Ozone-Depleting Substances 80
11.2. Estimated Current and Projected World Use of Potential
Ozone-Depleting Substances 80
A.I. Regression Estimates of Reporting Country CFC-11
and CFC-12 Production 89
A.2. Reported and Predicted Reporting Country CFC-11 and
CFC-12 Production 89
A.3. Regression Estimates of U.S. CFC-11 and CFC-12
Production 91
A.4. Reported and Predicted U.S. CFC-11 and CFC-12
Production 91
A.5. Regression Estimates of U.S. Methyl Chloroform and
Carbon Tetrachloride Production 92
A.6. Reported and Predicted U.S. Methyl Chloroform and
Carbon Tetrachloride Production 92
B.I. Comparison of Estimated 1985 CFC-11 Use 94
B.2. Comparison of Estimated 1985 CFC-12 Use 95
B.3. Comparison of Rand Estimates of 1976 and 1985
U.S. Nonaerosol CFC-11 and CFC-12 Use 96
B.4. Comparison of Rand and DuPont Estimates of 1976
U.S. Nonaerosol CFC-11 and CFC-12 Use 98
B.5. Estimated Shortfall in Rand 1976 Estimates of U.S.
CFC-11 and CFC-12 Use Compared with DuPont
Estimates 99
-------
I. INTRODUCTION
Release of several chlorofluorocarbons (CFCs) and related chemicals
to the atmosphere may reduce the concentration of ozone in the strato-
sphere. Almost all emissions of these chemicals result from human
activities. Depletion of stratospheric ozone would increase the amount
of ultraviolet radiation that penetrates to the earth's surface, poten-
tially causing increases in human skin cancers and adverse effects on
plants, animals, and marine life. In addition, when these substances
are released to the atmosphere they may reduce the radiant cooling of
the earth, thereby creating a "greenhouse effect," possibly warming the
earth's surface and changing its climate.1
Potential ozone-depleting chemicals are used in a wide range of
products. They are used to manufacture foam cushioning and insula-
tion products, as the refrigerant in refrigeration and air conditioning
systems, as solvents to clean electronic and other components in
manufacturing processes, and as aerosol propellants and fire
extinguishants. This report describes the major world uses of these
chemicals, estimates the current quantities used in each product area,
and projects future use to 2000.
Estimates of current use and projected future use are reported
separately for the United States, the other non-communist countries,
and the communist countries. This distinction is relevant because we
have much better information on U.S. than on foreign use of most of
these chemicals, and better information on non-communist than on
communist use. Nearly all of the production of CFC-11 and CFC-12
outside the communist countries is by companies that report their pro-
duction to the Chemical Manufacturers Association (CMA); production
in the communist countries is not reported. Consequently, we divide
the world into three regions for analysis: the United States, the "other
reporting countries," and the "communist countries" (including the
Soviet Union, Eastern Europe, China, and the other communist Asian
nations). The non-communist countries are jointly called the "report-
ing countries."2
!For more information on the atmospheric processes and possible adverse conse-
quences see Ramanthan et al. (1985) or National Academy of Sciences (1976, 1979, 1982,
1984).
2We treat annual production and use of the chemicals within regions as equivalent,
since inventories and net regional imports and exports are generally small. As described
in Appendix A, however, the difference between production and use in a single nation,
the United States for example, may be significant. Systematic data on imports and
-------
The chemicals discussed in this report are limited to those believed
to pose the greatest threat to stratospheric ozone, if such a threat
exists, as indicated by computer models that simulate atmospheric
chemistry. They include CFC-11, CFC-12, CFC-113, methyl chloro-
form, Halon 1211, Halon 1301, and carbon tetrachloride. Current con-
sumption estimates for the three world regions are summarized in
Table 1.1.
Carbon tetrachloride and methyl chloroform are produced in the
greatest quantities but are not considered the most important potential
ozone depleters. Carbon tetrachloride is used primarily as a chemical
intermediate in the production of CFC-11 and CFC-12, so the quanti-
ties that may be emitted are only a fraction of total production
estimated here.3 Although most methyl chloroform is used as a solvent
and eventually released to the atmosphere, it is relatively unstable in
the atmosphere and consequently more likely to decompose without
reaching the stratosphere.
CFC-11 and CFC-12 are considered the most important potential
ozone depleters because of their high production levels and relative
depletion potency. CFC-113 use is significantly smaller, but CFC-113
is thought to be almost as effective a potential depleter as the two
other CFCs.
Table 1.1
ESTIMATED 1985 WORLD USE OF POTENTIAL
OZONE-DEPLETING SUBSTANCES
(In thousands of mta)
Chemical
Other
United Reporting Communist
World States Countries Countries
CFC-11
CFC-12
CFC-113
Methyl chloroform
Carbon tetrachloride
Halon 1301
Halon 1211
341.5
443.7
163.2
544.6
1,029.0
10.8
10.8
75.0
135.0
73.2
270.0
280.0
5.4
2.7
225.0
230.0
85.0
187.6
590.0
5.4
8.1
41.5
78.7
5.0
87.0
159.0
0.0
0.0
"Metric tons.
exports of each chemical needed to estimate the size of the difference are apparently not
publicly available, however.
See Quinn et al. (1986) for a more explicit analysis of possible chemical emissions.
-------
The Halons are produced in only small quantities, about 11,000 mt
each worldwide. However, they contain bromine, which may be a much
more effective ozone depleter per atom than the chlorine contained in
the other substances. Although the Halons may not be important
potential ozone depleters at present, their use is expected to grow
rapidly and they may become more important in the future.
As shown by Table 1.1, the United States currently accounts for
about one-quarter to one-half of world use of these potential ozone
depleters. The U.S. share of specific applications of CFC-11 and CFC-
12 is generally similar, except for use in aerosols where, because of a
U.S. ban on CFCs in "nonessential" aerosol applications, the United
States accounts for only about 5 percent of world use. Use in the com-
munist countries is estimated to be less than 20 percent of the global
totals. Because of the limited data available on communist use, the
text focuses on use in the CMA reporting countries. Estimates of com-
munist use are derived in Sec. X.
An important feature of these chemicals is their chemical stability:
They do not readily react with others. This stability contributes to
their safety and usefulness in a variety of applications but also to their
potential threat to stratospheric ozone. Because of this stability most
of these chemicals are believed to survive in the atmosphere 50 to 100
years or longer, so their concentrations in the lower atmosphere may
remain high for many years after release. Only a small fraction of the
molecules of these chemicals in the lower atmosphere is transported to
the stratosphere; there it is decomposed by ultraviolet radiation and
the freed halogen atoms may catalytically react with the ozone
molecules. Consequently, if depletion occurs it may persist for decades,
even if emissions are terminated. The radiative effect of these chemi-
cals as greenhouse gases would also be persistent.
Differences in the time pattern of chemical emissions between prod-
ucts may be important. In some applications, such as aerosol propel-
lants, the chemicals are released directly to the atmosphere as the aero-
sol product is used. In other uses, such as home refrigerators, the
chemical is contained in a hermetically sealed system and not emitted,
except for minor leakage, until the product is disposed of. If disposed
of before the seals are broken it is theoretically possible to recover the
chemical and so prevent emission, although it may not be practical. In
cases like this the chemical constitutes a bank that may be released
sometime in the future, or possibly never released.
Other chemicals may also affect the concentration of stratospheric
ozone. Most of those that are suspected ozone depleters are simple
organic molecules containing either chlorine or bromine. Other gases,
including methane and oxides of nitrogen, may increase or decrease
-------
stratospheric ozone concentrations, depending on the concentrations of
other perturbants. Carbon dioxide emissions are believed to generally
increase ozone concentrations. These gases are being examined in
other research efforts funded by the Environmental Protection Agency
(EPA) and other agencies. However, none of the other potential ozone
depleters is thought to be as likely to pose a threat as the chemicals
discussed here. The other chemicals may be less important either
because they are produced in smaller quantities or because they are less
likely to reach the stratosphere. Perhaps the most widely used of these
less important chemicals is CFC-22, which is extensively used as the
refrigerant in air conditioners and other refrigeration equipment.
CURRENT APPLICATIONS
The seven chemicals we discuss are used in the following applica-
tions: aerosol dispensers, rigid foam insulation and related products,
flexible cushioning foams, refrigeration and air conditioning systems,
solvents, fire extinguishants, and other products.
CFC-11 and CFC-12 are used in diverse applications. In the CMA
reporting countries, CFC-11 is used almost entirely as an aerosol pro-
pellant and as a blowing agent in producing rigid and flexible foams.
As shown by Fig. 1.1, these three uses account for an estimated 90 per-
cent of reporting country use. Most of the remainder represents the
difference between the sum of our estimates of CFC-11 use in each
product area and total use, based on total production as reported by
CMA (1985) for the last several years (see Appendix A for details).
In the United States, CFC use in "nonessential aerosols" was
banned by the EPA effective in 1979 (the ban was announced in 1977).
As a result, aerosols account for only about 5 percent of U.S. CFC-11
use, whereas rigid foams account for about half. The manufacture of
flexible foam accounts for about one-fifth, similar to its share of world
use. The unallocated uses, about 18 percent, represent the difference
between the applications for which we estimate current use and 1985
domestic use estimated from U.S. International Trade Commission
(ITC) production data (see Appendix A). We believe that part of the
unallocated use of CFC-11 is in refrigeration and air conditioning (see
Appendix B). In addition, it includes storage, packaging, and transport
losses that may account for about 2 percent of use (see Wolf, 1980).
Reporting country and U.S. use of CFC-12 are shown in Fig. 1.2.
An estimated 32 percent of CFC-12 use in the reporting countries is as
an aerosol propellant, whereas 27 percent is used in the refrigeration
applications we analyze, including mobile air conditioning, retail food
-------
CMA Reporting Countries
(300,000 mt)
United States
(75,000 mt)
Unallocated
(8%)
Chillers (3%
Flexible molded
(4%)
Flexible
slabstock
(15%)
Rigid foam
(39%)
Aerosol
(31%)
Unallocated
(18%)
Chillers (6%)
Flexible
molded
(5%)
Flexible
slabstock
(15%)
Aerosol
(5%)
Rigid foam
(51%)
Fig. 1.1—Estimated CMA reporting country and U.S. use of CFC-11, by product
-------
cr>
CMA Reporting Countries
(365,000 mt)
United States
(135,000 mt)
Unallocated
(22%)
Miscellaneous
(7%)
Home
refrigerators
(3%)
Chillers (1%)-
Retail food
refrigeration (3%)
Mobile
air conditioning
(20%)
Aerosol
(32%)
Rigid
foam
(12%)
Unallocated
(31%)
Miscellaneous
(10%
Home refrigerators
(2%)
Aerosol
(4%)
Chillers (1%) Retail
food
refrigeration
(4%)
Rigid
foam
(11%)
Mobile
air conditioning
(37%)
Fig. 1.2—Estimated CMA reporting country and U.S. use of CFC-12, by product
-------
refrigeration, chillers (large commercial and industrial air conditioning
systems), and home refrigerators and freezers. The unallocated uses
total 22 percent.
In the United States, aerosols account for only about 4 percent of
use, about the same share as of CFC-11. The refrigeration applications
account for 44 percent, of which mobile air conditioning is by far the
largest component. The unallocated uses are about 31 percent of
estimated total domestic use. As discussed in Appendix B, we suspect
that a large part of the unexplained use of CFC-12 is in food refrigera-
tion applications, both in the United States and abroad. Again,
storage, packaging, and transport losses may account for about 2 per-
cent of total use.
The other chemicals are concentrated in fewer applications. Methyl
chloroform and CFC-113 are used almost exclusively as solvents.
CFC-113 is used largely in the electronics industry to "deflux" printed
circuit boards and clean plastic parts including semiconductors; methyl
chloroform is a general purpose solvent used in many types of metal
and other cleaning applications. A small amount of CFC-113 is also
used in specialty refrigeration applications.
Carbon tetrachloride is produced in larger quantities than any of the
other chemicals but most is transformed into CFC-11 or CFC-12.
Additional carbon tetrachloride is used as a solvent, as a grain fumi-
gant, and in the pharmaceutical industry, but chemical producers in
the United States have agreed to stop using it as an active ingredient
in pesticides by 1986.
Halon 1211 and Halon 1301 are relatively new and are produced in
small but growing quantities. They are used as fire extinguishants,
Halon 1301 primarily in total flooding systems used to protect com-
puter installations and other expensive equipment, Halon 1211 pri-
marily in hand-held extinguishers.
SUMMARY OF PROJECTED USE
Table 1.2 summarizes current and projected annual global use of the
seven potential ozone depleters we analyze. As described in Sec. II, our
projections include a base production level assuming continuation of
current and foreseeable trends and a range of uncertainty surrounding
the base projection. They assume no change in the perceived likeli-
hood of regulations on use or emissions of these chemicals because of
the threat of potential ozone depletion. Thus, they are meant to
characterize future use in the absence of additional regulations on
potential ozone depleters as such. The projections do account for the
-------
Table 1.2
ESTIMATED CURRENT AND PROJECTED WORLD USE
OF POTENTIAL OZONE DEPLETERS
(In thousands of mt)
2000 Use
Chemical
CFC-11
CFC-12
CFC-113
Methyl chloroform
Carbon tetrachloride
Halon 1301
Halon 1211
1985
Use
341.5
443.7
163.2
544.6
1029.0
10.8
10.8
Base
Use
560
620
420
840
1550
20
20
Annual
Base Growth
Rate (%)
3.3
2.3
6.5
3.0
2.8
4.4
4.2
Range of
Uncertainty
-140 to +180
-160 to +210
-130 to +180
-220 to +290
-390 to +510
-8 to +12
-8 to +12
possibility of other regulations, such as the U.S. ban on land disposal
of chlorinated solvents and regulations affecting substitute chemicals.
Table 1.2 characterizes the projections in terms of the projected base
level of use, the average annual growth rate necessary to achieve that
base level, and the range of uncertainty about the future level of use.
Recall that the chemicals differ in the degree to which they may reduce
the concentration of stratospheric ozone because of differences in emis-
sion patterns and chemical reactions in the atmosphere. Although car-
bon tetrachloride is produced in the largest quantities, because its dom-
inant use is as a chemical intermediate only a small fraction of current
production is released to the atmosphere (we estimate about 6 percent;
see Quinn et al., 1986). Similarly, methyl chloroform, despite its large
production volume, is not believed to be as important a potential ozone
depleter as the three CFCs, because much of it decomposes in the lower
atmosphere. In contrast, the Halons are thought to be very potent
ozone depleters and thus merit attention despite their low production
levels.
The seven chemicals may be divided into two groups on the basis of
their expected future growth rates. As shown in Table 1.2, the baseline
projections for CFC-11, CFC-12, methyl chloroform, and carbon tetra-
chloride involve growth at an average annual rate of about 3 percent.
In contrast, CFC-113 and the two Halons are projected to grow much
more rapidly: Their base projections entail about 4 to 6 percent
-------
average growth until 2000. This contrast reflects a fundamental differ-
ence between the two groups of chemicals: The first, slowly growing
group includes chemicals that have been produced for many years and
have found applications in a variety of uses. In contrast, the second
group is composed of "specialty" chemicals that are used in a narrower
set of applications but are making rapid inroads in these markets.
Table 1.2 indicates that absolute uncertainty about future produc-
tion is greatest for carbon tetrachloride and methyl chloroform,
because of their high production levels. Future CFC-12 production is
somewhat less certain than future CFC-11 production. The range for
CFC-113 is almost as broad as for CFC-11, in part because the
expected growth rate is higher. In contrast, the ranges for the Halons
are significantly smaller because of their low production levels.
Tables 1.3 and 1.4 report the projected use of CFC-11 and CFC-12
in each application.4 CFC-11 is used primarily as an aerosol propellant
Table 1.3
ESTIMATED CURRENT AND PROJECTED WORLD USE OF CFC-11
(In thousands of mt)
2000 Use
Applications in
Reporting Countries
Aerosol
Foam production
Rigid
Slabstock
Molded
Refrigeration and
air conditioning
Centrifugal chillers
Miscellaneous and
unallocated uses
Communist countries
Total world use
1985
Use
93.7
115.8
45.0
12.0
9.9
23.6
41.5
341.5
Annual
Base Base Growth Range of
Use Rate (%) Uncertainty
95
260
74
12
11
39
100
560
0.1
5.5
3.4
0.3
0.9
3.4
6.0
3.3
-37
-67
-26
-4.2
-3.2
-9.8
-30
-140
to
to
to
to
to
to
to
to
+59
+89
+39
+6.2
+4.3
+13
+42
+180
4The range of uncertainty for total use of the chemicals cannot be derived by adding
the ranges for each application. As explained in Sec. II the ranges represent subjective
80 percent credibility regions. The region for total use derived by adding the regions for
each application would correspond to a much higher credibility level. See Camm and
-------
10
Table 1.4
ESTIMATED CURRENT AND PROJECTED WORLD USE OF CFC-12
(In thousands of mt)
2000 Use
Applications in
Reporting Countries
Aerosol
Foam production
Rigid
Refrigeration and
air conditioning
Mobile air cond.
Retail food refrig.
Home refrigeration
Centrifugal chillers
Reciprocating chill.
Miscellaneous and
unallocated uses
Communist countries
Total world use
1985
Use
115.6
42.8
73.4
9.7
10.2
3.7
1.3
108.3
78.7
443.7
Annual
Base Base Growth Range of
Use Rate (%) Uncertainty
120
76
110
6.4
22
4.3
0.5
140
160
620
0.1
3.9
3.0
-2.7
5.2
1.0
-5.7
1.7
5.0
2.3
-46
-31
-34
-3.5
-7.6
-1.2
-0.3
-38
-53
-160
to
to
to
to
to
to
to
to
to
to
+72
+51
+48
+7.3
+ 11
+1.6
+0.5
+51
+75
+210
and as a blowing agent in foam manufacturing. Growth trends in these
applications are very dissimilar, however. Use in aerosols in the
reporting countries is projected to remain approximately constant. In
contrast, use in rigid foams is projected to grow rapidly, with a base
projected growth rate above 5 percent. Largely because of this higher
expected growth, the absolute uncertainty about CFC-11 use in rigid
foam is substantially larger than the uncertainty about aerosol use.
Aerosol use of CFC-12 is also expected to remain about constant, as
shown in Table 1.4. The next largest applications are mobile air condi-
tioning and rigid foam production. Growth in these areas is expected
to be strong, although the corresponding ranges of uncertainty include
nearly constant use in each area.
Hammitt (1986) for more detail on our method for calculating the ranges. Because of
approximations in that method the projected total base use does not equal the sum of the
base uses in each application.
-------
11
ORGANIZATION OF THE REPORT
Section II describes the methodology used to project future chemical
use in each product area and to characterize the uncertainty associated
with each projection. The following sections describe the diverse prod-
ucts in which the chemicals are used in more detail and report
estimated current U.S. and other reporting country use and projected
use to 2000.
The products are described in roughly decreasing order of their
potential threat to stratospheric ozone, as measured by the quantities
of potential ozone-depleting chemicals used, adjusted for their approxi-
mate estimated ozone-depletion potency. Within product areas, indi-
vidual products are also discussed in decreasing order of importance.
Sections III through VII describe reporting country use of CFC-11 and
CFC-12 in aerosol products, in manufacturing rigid foams for insula-
tion and other applications, flexible foams for cushioning, as a refriger-
ant in refrigeration and air conditioning systems, and in miscellaneous
uses.
Section VIII describes solvent and other uses of CFC-113, methyl
chloroform, and carbon tetrachloride (which is used primarily as a
chemical intermediate in CFC-11 and CFC-12 production). Section IX
describes fire extinguishant applications of Halon 1301 and Halon
1211. Section X estimates current and future use of all seven of the
potential ozone-depleting substances we consider in the communist
countries. Section XI summarizes the analyzed product applications
and compares them with total use reported by the CMA and the U.S.
ITC.
Supporting material is included in three appendices. Appendix A
documents our method for estimating 1985 total use of the chemicals
for which reported production data are available. Appendix B com-
pares the sum of the estimates of current use in each major application
of CFC-11 and CFC-12 to the total production reported by other
sources. It then attempts to determine the source of the differences by
comparing the estimates in this report with other sources. Appendix C
elaborates on the method for characterizing the uncertainty about pro-
jected use described in Section II.
-------
II. METHODOLOGY
The descriptions of product use and manufacturing processes are
based on intensive interaction with industry that began in the 1970s.
Over this period a large number of industrial companies and trade
organizations cooperated with Rand by providing basic data, current
estimates, and projections. More detailed descriptions of nonaerosol
CFC use in the United States can be found in Palmer et al. (1980).
To date, most of our information has come from U.S. manufacturers
and trade associations. Although we are actively pursuing information
outside the United States, our current data are necessarily more lim-
ited.
Projections of future chemical use in each application are based on
analysis of historical trends, industry projections, and comments. In
many cases we project product use to remain a constant share of the
U.S. or other national or regional economy—that is, to grow at a rate
equal to the corresponding Gross National Product (GNP).
As we obtain additional information our estimates of current and
likely future use will inevitably change. In a project of this nature
there can never be a "final word," only a most recent estimate. This
report summarizes our information as of early 1986.
CHARACTERIZATION OF THE UNCERTAINTY ABOUT
PROJECTED CHEMICAL USE
Projecting the outcome of future events is always a hazardous under-
taking. In this report we project a range of possible chemical uses in a
number of applications. For expositional clarity we characterize each
range by a base projection and a range of uncertainty for the year 2000.
The base projection is intended to characterize the middle of the distri-
bution of credible future outcomes. The range of uncertainty is
described by two numbers that indicate the approximate upper and
lower ends of what appears to us to be the range of reasonable out-
comes, relative to the base projection. More precisely, we think that
the probability that the future outcome will fall within the correspond-
ing range is 0.8: Thus the ranges represent subjective 80 percent credi-
bility intervals.
An alternative style for presenting our projections would be to report
the limits of the range of uncertainty directly, without a central base
projection. However, this style of exposition would assign too much
12
-------
13
legitimacy to the bounds of the range, which are inherently fuzzy. It
would also make the procedure for calculating the projections and the
assumptions used less transparent.
Although we report a range of uncertainty for chemical use only for
2000, projected use in earlier years should also be characterized by a
range. In most situations the range should be smaller for years closer
to the present. If required, we suggest that the appropriate ranges for
earlier years be calculated by assuming that the annual growth rates
are normally distributed, calculating the corresponding means and
standard deviations implied by the range for 2000, and from these cal-
culating the desired credibility intervals.1
The base projections that characterize the center of the range of
projected use are based on current and anticipated trends in chemical
use in each application. They assume no unanticipated shocks, such as
changes in the perceived likelihood that use will be regulated because
of concern about possible ozone depletion, or unexpected changes in
relative prices. Moreover, we do not attempt to project short-term
oscillations in use associated with business cycles but instead try to
characterize the basic trends.
The limits of the ranges of reasonable uncertainty are derived from
explicit consideration of three levels of uncertainty:
• Uncertainty about the level of general economic activity,
described by the GNP,
• Uncertainty about final product use conditional on GNP, and
• Uncertainty about the use of the chemical per unit of final
product, conditional on final product use. Chemical use per
unit of product may change because of changes in the relative
costs of alternative manufacturing technologies resulting from
technological innovation, regulation, or changes in real input
prices.
We assess ranges of reasonable uncertainty for each of the three lev-
els of uncertainty and combine them to produce an overall range using
the procedure described in Appendix C.2 The range of uncertainty
about GNP growth is common to all of the product applications. It is
described in the following subsection. The other ranges are based on
Camm and Hammitt (1986).
2We use log-normal distributions to approximate our subjective probability distribu-
tions corresponding to each level of uncertainty. Although the uae of log-normal distri-
butions imposes some restrictions on the character of the uncertainty that can be
represented, they provide close approximations to our subjective distributions in most
cases. The assumption that chemical use is distributed log-normally is equivalent to the
assumption of normally distributed growth rates.
-------
14
our knowledge of specific product areas and our beliefs about what
changes might reasonably occur within the next 15 years.
Because part of the uncertainty about use in each application—that
due to uncertainty about general economic growth—is common to all
applications of a given chemical, derivation of the range of uncertainty
for total use of a chemical requires accounting for this correlation
among uses. The reported ranges of total CFC-11 and CFC-12 use
account for this correlation using the procedure described in Camm
and Hammitt (1986).
The chemical use projections do not explicitly account for uncer-
tainty about current use. As noted in Sec. I, our "bottom up" estimates
of use in each major application do not account for all of the CFC-11
and CFC-12 currently used. There may be other significant uses we
have not analyzed, or our estimates for current use in some of the
applications may be too low. It would be possible to develop a subjec-
tive range of uncertainty for current use, both in applications and in
total, and to incorporate this uncertainty with the uncertainty about
future growth. However, we have not yet taken this step. If future
information reveals our estimates of current use in a specific applica-
tion to be too low we would recommend revising the corresponding
range of projected use upward.
Our ranges of uncertainty are reported in the discussion of each
product area. Readers who find our intervals inappropriate are invited
to supply their own and see how it affects the final range for total
chemical use.
PROJECTED GROWTH IN GNP
In many of the baseline projections final products are assumed to
remain a constant share of a nation's economy, that is, to grow at the
same rate as the GNP. The base GNP growth rates we assume are
reported in Table 2.1. With one exception, the rates are based on
those prepared by Edmonds et al. (1984) for the U.S. Department of
Energy. These projections are a revision of the rates used in our ear-
lier work projecting long-term growth (reported in Quinn et al., 1986).
They are based on analysis of likely population and labor productivity
trends and, compared with other published projections, are low to
moderate.3
The one difference between our rates and the Edmonds et al. (1984)
projections is for the United States. Edmonds et al. (1984) project a
3See Quinn et al. (1986), Appendix B.
-------
15
Table 2.1
PROJECTED BASE GNP GROWTH RATES
(In percent per year)
Region 1985-2000
United States 3.0
Western Europe and Canada 2.8
Japan, Australia, and New Zealand 4.0
USSR and Eastern Europe 2.4
Centrally Planned Asia 4.5
Mideast 6.6
Africa 3.7
Latin America 5.3
South and East Asia 5.1
Other reporting countries 3.5
Communist countries 3.0
World 3.3
base growth rate of 2.7 percent per year. For the period 1985 to 1990
we have increased the base rate to 3.5 percent, consistent with
Congressional Budget Office (1985) projections. This estimate is
slightly lower than the prominent alternative of 4.0 percent suggested
by the Council of Economic Advisors (1985). After 1990 we adopt the
Edmonds et al. (1984) projection of 2.7 percent per year. The resulting
average rate over the 15 year period is 3.0 percent.
For most product areas we do not disaggregate use in the CMA
reporting countries outside the United States. Rather we use an aggre-
gate growth rate of 3.5 percent for other reporting country GNP
growth, which is the appropriate aggregate of the regional rates in
Table 2.1.
As described above, we assessed ranges of uncertainty around the
base GNP growth rates. For the United States our intervals include
growth rates 1.5 percent higher or lower than the base projections.
Thus, our intervals extend from 1.5 to 4.5 percent per year over the
period. These intervals are based on the observation that, since World
War II, U.S. growth has averaged between about 1.5 and 4.5 percent
per year over periods of about a decade. For the other CMA reporting
countries we use the same range of uncertainty as for the United
States, plus or minus 1.5 percent from the base, or 2.0 to 5.0 percent
per year. Although uncertainty about growth in some parts of the
world is greater than for the United States, variations in different
-------
16
regions will tend to partially offset one another so the range of uncer-
tainty about the growth of the total is less than the range for an indi-
vidual country. For the communist countries we use a wider range,
plus or minus 2.0 percent (a range of 1.0 to 5.0 percent averaged over
the period). We feel that a wider range is appropriate for these coun-
tries because we have less information about their likely growth and
because the communist countries constitute a smaller region than the
other reporting countries, and hence variations within the region are
less likely to offset each other. The ranges of uncertainty implied by
these ranges of possible growth rates are between 0.80 and 1.25 times
the base projected GNP level in 2000 for both the United States and
the other reporting countries, and between 0.74 and 1.35 times the base
projected level in the communist countries.
-------
III. AEROSOL PROPELLANTS
CFC-ll and CFC-12 are used extensively as propellants in a wide
variety of aerosol products. Many of these are personal care products,
such as deodorant, perfume, and hairspray; others include household
products such as insecticides and drain cleaners, and household and
industrial products, such as spray paints and lubricants. Substitute
propellants include hydrocarbons and carbon dioxide, and in some
applications aerosol cans can be replaced by pump dispensers or other
nonpressurized applicators. CFCs are more expensive than the alterna-
tive aerosol propellants but they are not flammable like hydrocarbon
propellants, and some of their attributes make them technically more
attractive in limited applications.
Because the propellant is emitted as the aerosol is discharged emis-
sions are typically prompt. Before 1975 aerosol use was the largest sin-
gle source of worldwide CFC emissions.
Historical CFC use in aerosols is reported in Table 3.1 and Fig. 3.1.
In 1978 the EPA banned use of CFC propellants in all aerosols except
a limited list of "essential uses" (including military applications).
Under the ban United States use fell about 95 percent from its peak
(1973) level.1 Subsequently the European Economic Community nego-
tiated a 30 percent reduction in use. Current EEC use is about 45 per-
cent below use in the late 1970s, well below the target reduction, as the
voluntary reductions have apparently stimulated increased substitution
of less expensive hydrocarbon and carbon dioxide propellants.
United States use has remained constant at about 9500 mt since the
ban and no significant change is expected. We project constant use
with an uncertainty range of 78 to 128 percent of the projected value in
2000. This range is composed of uncertainty in the level of aerosol use,
conditional on the GNP, of 90 to 111 percent of the projected base
level, combined with the uncertainty range for the U.S. GNP.
Use in the European Economic Community has declined steadily
since the mid 1970s as manufacturers have substituted other propel-
lants. The substitution of pump dispensers for aerosol cans is
apparently not occurring to the extent it did in the United States.
!United States CFC use in aerosols fell substantially between 1973 and the announce-
ment of the ban in 1977. This decline is probably related to several factors including the
Arab oil embargo, anticipation of the regulation, and the discovery that additional aero-
sol products could be successfully marketed with non-CFC propellants. (Some aerosols
were traditionally propelled without CFCs; for example, spray paint has usually employed
hydrocarbons.)
17
-------
18
Fundamental uncertainty exists as to how much further CFC use will
decline there. It is possible that all of the decline that will occur in the
absence of regulation has all ready occurred. Alternatively, reductions
in use could parallel those in the United States. However, the level of
peak use is uncertain. By analogy to the United States, we assume
that the peak occurred in 1973 and was 50 percent higher than the
1976 level, that is, about 265,000 mt. If this estimate is correct, 1983
use is about 37 percent of estimated peak (1973) use.
We project EEC use to fall to about 50,000 mt or almost 20 percent
of the assumed peak by 2000. As an upper bound on the range of
uncertainty we use the current level of about 100,000 mt, since we do
not expect CFC use in aerosols to grow in the EEC. The lower bound
is 25,000 mt or almost 10 percent of the assumed peak. This amount is
about twice the level of U.S. use (5 percent of peak use), which seems
reasonable assuming no further regulation in the EEC.
Table 3.1
ESTIMATED HISTORICAL USE OF CFC-11 AND CFC-12 IN AEROSOLS
(In mt)
Year
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
United
States8
156,400
161,400
185,000
212,700
198,600
162,700
138,600
79,500
40,000
9,500
9,500
9,500
9,500
9,500
9,500
EEC"
na
na
na
na
na
na
176,900
162,600
150,400
136,600
126,400
116,100
111,700
98,800
na
Japan0
na
na
na
na
14,400
14,000
15,500
14,000
13,300
12,900
13,000
12,500
12,000
11,800
11,800
Other
Reporting
Countriesd
na
na
na
na
na
na
101,200
94,700
103,500
95,600
94,200
83,500
75,300
91,900
na
Total
Reporting
Countries'
na
na
na
na
na
na
432,300
350,800
307,200
254,600
243,300
221,600
208,500
212,000
218,800
NOTE: na = not available.
•Wolf (1980).
bEuropean Fluorocarbon Technical Committee data as supplied by EEC.
'Ministry of International Trade and Industry data as supplied by Japan
Flon Gas Association (private communication, 1985).
dTotal reporting countries less United States, EEC, Japan.
'CMA (1985).
-------
450
400
350
NOTE: Earlier reporting country use not available
«? 300
E
8
25°
200
V)
16
I 150
100
50
J L
Reporting Countries
(total CFC-11 and CFC-12)
U.S. (total CFC-11 and CFC-12)
I I I I I I 1 . I 1 1 1 1 1
J 1 L
1970
1975
1980
1985
Year
1990
1995
2000
Fig. 3.1—Estimated historical and projected U.S. use of CFC-11 and CFC-12 in aerosols
-------
20
Japanese use has declined slowly, in part because of substitution of a
blend of CFCs and hydrocarbons. Since Japanese law prohibits using
pure hydrocarbons as aerosol propellants (because of fire hazards) any
substitution to other propellants is presumably to carbon dioxide; sub-
stitution from aerosols to pumps may also be occurring. As these alter-
natives are technically inferior in many uses, we use a base projection
for Japanese CFC use equal to current production of about 12,000 mt
with a range of uncertainty from three-quarters to one and one-third
times that amount.
Further growth is likely in the industrializing nations, particularly in
East Asia. These countries are undergoing rapid economic growth and
are not party to the international agreements restricting CFC use in
aerosols. Our base projection assumes modest growth of 2.5 percent,
one-half the expected rate of GNP growth. We believe growth could
reasonably range from one-half to twice as fast.
The base projections are summarized in Table 3.2 and Fig. 3.1.
They assume that the proportions of aerosol production accounted for
by each CFC remain constant at the current reporting country propor-
tions: about 45 percent CFC-11 and 55 percent CFC-12. Note that the
projected declines in the EEC almost exactly offset projected growth in
the other reporting countries: Baseline projected use initially declines
then rises as the developing countries account for a larger share of the
total. Combined with the anticipated constant U.S. use, the offsetting
trends in the EEC and other reporting countries produce a nearly con-
stant base projection for reporting country aerosol use. As noted
Table 3.2
ESTIMATED CURRENT AND PROJECTED USE OF
CFC-11 AND CFC-12 IN AEROSOLS
(In mt)
Other Reporting All Reporting
United States Countries Countries
Year CFC-11 CFC-12 CFC-11 CFC-12 CFC-11 CFC-12
1985
1990
1995
2000
3,800
3,800
3,800
3,800
5,700
5,700
5,700
5,700
89,900
88,100
88,500
90,800
109,900
107,700
108,200
111,000
93,700
91,900
92,300
94,600
115,600
113,400
113,900
116,700
-------
21
above, the range of uncertainty for U.S. use is between 78 and 128 per-
cent of projected base use in 2000. Outside the United States the
range is from about 60 to 165 percent of projected base use in 2000.2
2This calculation assumes that use is correlated across regions, with a correlation
coefficient of 0.75.
-------
IV. RIGID FOAM
CFC-ll and CFC-12 have been used to produce rigid or "closed-cell"
foams since the early 1960s. These products are widely used for insula-
tion and packaging.
CFC-ll and CFC-12 are used as blowing agents to form the holes or
cells when the foam is manufactured. In closed-cell foams the CFCs
remain trapped inside, diffusing only slowly to the atmosphere.
Because of the low thermal conductivity of the CFCs, these foams are
about twice as effective insulators as equally thick nonfoam alterna-
tives. Foam sheets or boards are sometimes clad with aluminum,
paper, or asphalt sheathing to slow the diffusion of air into the foam
and thus slow the degradation of its insulating capacity.1
There are two major types of closed-cell foams: rigid urethane or
isocyanurate foam, usually produced using CFC-ll, and nonurethane
foam, usually produced with CFC-12. Rigid uretharie foam accounts
for the second largest share of world CFC-ll use, after aerosols, and by
far the largest share of U.S. CFC-ll use. Rigid urethane foams are
better insulators than the nonurethane foams, and over 90 percent of
these foams are used as insulation in residential, commercial, and
industrial construction, in home and commercial refrigerators, and in
refrigerated trucks and railroad cars. Insulating foam is made into
sheets or boards that are installed during construction, sprayed directly
onto tanks, pipes, and other structures, or foamed in place, as inside
the walls of a refrigerator. Noninsulation applications include flotation
devices, marine buoys and use as a packaging material to protect valu-
able and delicate objects during shipment.
The most important CFC-consuming nonurethane foams—extruded
polystyrene (PS) sheet and extruded PS board—use smaller quantities
of CFC-12. PS sheet products include meat trays, egg cartons, plates,
and other food containers. Extruded PS board was introduced by Dow
Chemical Company during World War II under the trade name Styro-
foam. It is used primarily as insulation and competes directly with
rigid urethane foam in the residential and commercial wall and roof
sheathing market.
Insulating capacity may be degraded by changes in the amount and composition of
gas retained in the foam, which apparently depend on complicated diffusion and dissolu-
tion processes. See Mooz et al. (1982).
22
-------
23
RIGID URETHANE FOAM
Rigid urethane foam is used primarily for insulation in construction,
refrigeration, and transportation applications. Table 4.1 presents his-
torical U.S. foam use in these markets.
We estimate that each kilogram of foam requires an average of
about 130 grams of CFC-11 to produce.2 From the quantities of CFC-11
required per unit of foam we estimate CFC-11 use by application,
reported in Table 4.2. As shown, the construction market currently
accounts for about two-thirds of CFC-11 use, refrigeration for one-fifth,
and transportation for one-ninth. Estimated reporting country use of
CFC-11 in rigid urethane foams is presented in Table 4.3. United
States use accounts for about one-third of the total.
The markets for rigid urethane insulating foam are closely related to
the construction industry and tend to follow similar cycles. However,
growth in most insulation markets has been strong and is expected to
continue. Our base projections for U.S. use are reported in Table 4.4.
In the United States and Western Europe, growth in rigid foam use
in construction averaged 12 percent annually between 1976 and 1981
(before the major recession in the early 1980s; see Table 4.3), whereas
U.S. growth has historically averaged 5 to 8 percent (see Table 4.2).
Table 4.1
ESTIMATED HISTORICAL U.S. PRODUCTION OF
RIGID URETHANE FOAM
(In mt)
Year
1978
1979
1980
1981
1982
1983
1984
Construction
131,000
154,000
142,000
163,000
152,000
163,000
177,000
Insulation
Refrigeration
45,000
48,000
48,000
50,000
48,000
50,000
53,000
Transportation
21,000
22,000
18,000
22,000
16,000
21,000
32,000
Other
18,000
20,000
18,000
11,000
10,000
11,000
12,000
Total CFC-
Blown
215,000
244,000
226,000
246,000
226,000
245,000
274,000
SOURCE: Modern Plastics (various issues).
Specifically, we estimate the following average quantities of CFC-11 to produce one
kilogram of foam for each application: 126 (construction), 140 (refrigeration), 136 (trans-
portation), and 137 (other applications). See Palmer et al. (1980) Tables III.C.2, F.2, and
F.3.
-------
24
Table 4.2
ESTIMATED HISTORICAL U.S. USE OF CFC-11 IN RIGID URETHANE FOAM
(In mt)
Year
1978
1979
1980
1981
1982
1983
1984
Construction
16,500
19,400
17,900
20,500
19,200
20,500
22,300
Refrigeration
6,300
6,700
6,700
7,000
6,700
7,000
7,400
Transportation
2,900
3,000
2,400
3,000
2,200
2,900
4,400
Other
2,500
2,700
2,500
1,500
1,400
1,500
1,600
Total
28,200
31,900
29,500
32,000
29,400
31,900
35,700
Table 4.3
ESTIMATED HISTORICAL REPORTING COUNTRY USE
OF CFC-11 IN RIGID URETHANE FOAM
(In mt)
Year
1976
1977
1978
1979
1980
1981
1982
1983
1984
United
States8
18,500
21,800
28,200
31,900
29,500
32,000
29,400
31,900
35,700
EEC"
20,100
21,500
27,300
29,000
33,400
35,200
na
na
na
Other
Reporting
Countries0
13,500
21,900
10,600
19,200
21,200
30,500
na
na
na
Total
Reporting
Countries'1
52,100
65,200
66,100
80,100
84,100
97,700
94,900
na
na
NOTE: na = not available.
"From Table 4.2.
bBased on industry sources.
Calculated as total reporting countries less United States
and EEC.
dCMA reports.
-------
25
Table 4.4
ESTIMATED CURRENT AND PROJECTED U.S. USE
OF CFC-11 IN RIGID URETHANE FOAM
(In mt)
Year
1985
1990
1995
2000
Construction
24,100
35,400
45,200
57,600
Refrigeration
7,600
8,700
9,800
11,100
Transportation
4,800
7,800
9,900
12,700
Other
1,800
2,000
2,000
2,000
Total
38,300
53,900
66,900
83,400
The base projections for U.S. use in Table 4.4 assume that high growth
averaging 8 percent will persist through 1990, followed by 5 percent
average growth to the end of the century.
The refrigeration market is relatively mature and historic U.S.
growth has been similar to the GNP. The base projections assume
growth slightly slower than the GNP, 2.7 percent to 1990 and 2.5 per-
cent thereafter.
The transportation market accounts for only a small share of use
but is growing most rapidly, nearly doubling in the first half of the
1980s. The current 15 percent growth rate is unlikely to be sustained,
however. Our base projections assume that the rate will average 10
percent for the remainder of the decade and 5 percent over the next
decade.
The noninsulation rigid urethane markets declined significantly dur-
ing the early 1980s and never recovered to mid 1970s levels. Here, we
assume that they recover only slightly further, leveling off at 200 mt of
CFC use by 1990.
Table 4.5 presents base projections for reporting country use of
CFC-11 in rigid foams. Rigid urethane foam use in Western Europe
has historically been similar to U.S. use and the projections assume
similar growth. Other reporting countries have apparently experienced
more rapid growth (according to CMA reports), which is reasonable,
since they are undergoing more rapid economic development. The base
projections assume that the entire rigid urethane markets in these
countries will grow at the same rate as the rapidly growing construc-
tion markets in the United States and Europe, an average of 8 percent
through 1990 and 5 percent thereafter.
To estimate the range of uncertainty about the base projections we
start with the uncertainty about the size of final product markets con-
ditional on GNP. Because these are well established, at least in the
developed economies, we believe that uncertainty is similar to the
-------
26
Table 4.5
ESTIMATED CURRENT AND PROJECTED REPORTING
COUNTRY USE OF CFC-11 IN RIGID URETHANE FOAM
(In mt)
Year
1985
1990
1995
2000
United
States
38,300
53,900
66,900
83,400
Other
Reporting
Countries
77,500
111,100
139,600
175,900
Total
Reporting
Countries
115,800
165,000
206,500
259,300
uncertainty about general economic growth, between 0.8 and 1.25 times
the projected level. Additional uncertainty concerns the share of the
final product markets that will be served by CFC-11, since there exist a
number of other insulating products, such as fiberglass, extruded poly-
styrene board, and particle board, that are competitive in some applica-
tions. In contrast, the use of CFC-11-blown foam seems secure in
applications where the foam is sprayed or poured in place. Conse-
quently, we doubt that rigid urethane foams will lose more than 20 per-
cent, or gain more than 25 percent, of their current share. Combining
these ranges produces an overall range of uncertainty of 73 to 137 per-
cent of projected base use.
NONURETHANE FOAM
The major nonurethane products are extruded polystyrene board,
used for commercial and residential insulation, and extruded polysty-
rene sheet, used in egg cartons and food service trays. Jointly, they
account for about two-thirds of the CFC-12 used in nonurethane foam
products. Table 4.6 presents historical U.S. production of these foams
and estimated CFC-12 use. CFC-12 use per unit of nonurethane foam
is about 95 grams per kilogram of foam, smaller than CFC-11 use in
rigid urethanes.
CFC-12 is the primary blowing agent for extruded polystyrene sheet,
but use declined during the late 1970s and early 1980s. Palmer et al.
(1980) estimate CFC-12 use in 1976 assuming that 64 percent of poly-
styrene sheet is CFC-blown (the remainder uses pentane) and that
-------
27
Table 4.6
ESTIMATED HISTORICAL U.S. PRODUCTION OF
NONURETHANE FOAM AND USE OF CFC-12
(In mt)
Extruded PS Board Extruded PS Sheet
Year Foam CFC-12 Foam CFC-12
1976
1977
1978
1979
1980
1981
1982
1983
1984
na
na
na
31,000
27,000
32,000
30,000
34,000
39,000
1,900
2,300
2,500
2,900
2,600
3,000
2,900
3,200
3,700
na
na
na
151,000
147,000
140,000
146,000
153,000
184,000
5,900
7,300
7,800
7,600
7,400
7,000
6,600
6,500
7,200
SOURCES: Foam production from Modern Plas-
tics (various issues). CFC-12 use before 1979 from
Palmer et al. (1980), after 1979 calculated from foam
production (see text).
NOTE: na = not available.
CFC-blown foam contains 7.8 percent CFC-12 by weight. During
recent years, however, a mixed blowing agent combining CFC-12 and
carbon dioxide has reduced the CFC requirement by about 25 percent.
By 1984, we calculate CFC-12 use in polystyrene sheet assuming that
two-thirds of the foam uses 5.85 percent CFC-12 by weight.
Other nonurethane foams including expanded polystyrene foam
(used for drinking cups) and the polyolefins are manufactured in large
quantities but require only small amounts of CFC blowing agents. In
the absence of direct data on CFC use in these products we assume
that use grows at the same rate as in the extruded polystyrene foams.
Table 4.7 provides a base projection for U.S. use of CFC-12 in rigid
foams. For extruded PS board, we assume growth comparable to that
of rigid polyurethane insulation with which it competes. The base pro-
jections for use in polystyrene sheet and in the other nonurethane
category grow with GNP.
Table 4.8 presents base projections for reporting country use of
CFC-12 in nonurethane foam. Our base projections have the same
growth rates outside the United States as within. Since about half of
CFC-12 use in nonurethane foams is for extruded polystyrene sheet,
which faces a variety of competing products such as cardboard egg
-------
28
Table 4.7
ESTIMATED CURRENT AND PROJECTED U.S. USE
OF CFC-12 IN NONURETHANE FOAM
(In mt)
Extruded Polystyrene
Year
1985
1990
1995
2000
Board
4,000
5,900
7,500
9,600
Sheet
7,500
8,900
10,200
11,600
Total
11,500
14,800
17,700
21,200
Other
Nonure-
thanes
3,400
4,000
4,600
5,300
Total
Nonure-
thanes
14,900
18,800
22,300
26,500
cartons and food serving trays, and because some of these products can
also be made with pentane, we believe that there is significant uncer-
tainty about CFC-12 use conditional on the size of these packaging
markets. We suggest a range of uncertainty for CFC-12 use in
nonurethane foam from 60 percent of the projected level, if CFC-12-
blown polystyrene sheet were to lose 80 percent of its market, to 167
percent of the projected level if it were to displace many of its competi-
tors. Combining these with the range of uncertainty about GNP
growth produces a range of uncertainty of 57 to 172 percent of the base
projections.
Figures 4.1 and 4.2 summarize historical and base projected use of
CFC-11 and CFC-12 in rigid foams, in the United States and all
reporting countries, respectively. Note that the rigid urethanes account
for the majority of CFC use in rigid foam.
Table 4.8
ESTIMATED CURRENT AND PROJECTED REPORTING
COUNTRY USE OF CFC-12 IN NONURETHANE FOAM
(In mt)
Year
1985
1990
1995
2000
United
States
14,900
18,800
22,300
26,500
Other
Reporting
Countries
27,900
35,200
41,800
49,600
Total
Reporting
Countries
42,800
54,000
64,100
76,100
-------
Annual use (thousands of mt)
oo
r*
O
<-l
i—i.
O
05
t—'
05
3
a-
o
d
CA3
C
O
O
O
05
O
*J
O
i
t— '
to
cq
»•* •
a-
B5
05
3
-i O
co
-vl
o>
co
o
Ul
o
en
o
-M
o
oo
o
co
o
o
o
co
oo
o
CO
oo
CD
» s
CD
CD
co
N>
CD
CO
en
ro
o
o
o
-------
400
NOTE: Historical CFC-12 use not available
1976
1980
1996
2000
CO
o
Fig. 4.2—Estimated historical and projected reporting country use of CFC-11 and CFC-12 in rigid foam
-------
V. FLEXIBLE FOAM
CFC-ll is widely used to produce flexible foams, which first came
into use in the 1950s. Because these foams are resilient, durable, and
relatively inexpensive they are now widely used in furniture, bedding,
carpet underlay, automobile seats and dashboards, and packaging appli-
cations.
There are two major processes for producing flexible foams. Most
foam is manufactured in the form of slabstock, which is a piece of foam
typically about six feet wide, four feet high, and from six to 200 feet
long. Final shapes are cut from this large slab. The second process
molds foam into its final shape, which is often rounded or has other
features that would make it difficult or costly to carve from slabstock.
Molded foam is used almost exclusively for automobile seats and seat-
backs.
The important characteristics of flexible foam are imparted by blow-
ing agents that form the holes or cells in the foam and give it flexibil-
ity. All flexible foams use carbon dioxide as the primary blowing
agent. To produce softer (less dense) foams requires augmenting the
carbon dioxide with an auxiliary blowing agent, either CFC-ll or
methylene chloride. Since the carbon dioxide is produced by reacting
water with TDI (toluene diisocyanate), foams produced without aux-
iliary blowing agents are called water-blown.
Flexible foams are often called "open-cell," since their cells are open
to the atmosphere. As a result, CFC emissions are prompt: The blow-
ing agent escapes from the foam within a matter of hours or days after
its manufacture.
SLABSTOCK FOAM
Table 5.1 presents estimated U.S. production of slabstock foam and
CFC-ll used. Foam production is reported in Mobay (1982-1985).
CFC-ll use is based on estimated use of 50 grams per kilogram of
CFC-blown slabstock foam. CFC-ll use peaked in the late 1970s,
declined through 1982, and is currently increasing. The decline was
primarily due to reduced foam production during the recession, but it
also reflects reduced use of CFC-ll per unit of foam because of substi-
tution to the less expensive methylene chloride. Currently, methylene
chloride and CFC-ll each account for about half of the auxiliary blow-
31
-------
32
Table 5.1
ESTIMATED HISTORICAL U.S.
PRODUCTION OF SLABSTOCK
FOAM AND USE OF CFC-12
(In mt)
Year
1976
1977
1978
1979
1980
1981
1982
1983
1984
Foam
380,000
439,000
472,000
473,000
425,000
409,000
369,000
414,000
448,000
CFC-11
11,400
12,900
13,600
13,300
11,700
11,000
9,700
10,600
11,200
ing agent used.1 Further substitution to methylene chloride is not
expected because the cost savings will not offset the necessary invest-
ment in learning to produce sufficiently high quality foam. Future use
of methylene chloride may even decline because of concerns about its
potential carcinogenicity and possible regulation. A new alternative
blowing process has been recently patented. According to its Belgian
developers, the new process is less expensive, can be used to produce
all densities of foam, and uses no auxiliary blowing agent (CMR,
February 28, 1983). If these claims prove correct, adoption of this new
process could lead to substantial reductions in CFC-11 use.
Industry sources expect slabstock foam production to grow at the
same rate as the GNP (Mobay, 1985). Table 5.2 reports our base pro-
jections for U.S. production of slabstock foam and CFC-11 use, assum-
ing that CFC-11 use per unit of foam remains at its current level. We
suggest that these projections are subject to the following range of
uncertainty. First, the uncertainty concerning GNP growth implies
that total use could be about 20 percent smaller or 25 percent greater
than the base projection by 2000. We treat the uncertainty about the
growth of the final product markets, conditional on a given level of
GNP, as negligible. The level of CFC use, conditional on the size of
the final product markets, is subject to substantial uncertainty, how-
ever. Considerable substitution away from CFC-11 to either methylene
CFC-11 use averages 25 grams per kilogram of slabstock foam.
-------
33
Table 5.2
ESTIMATED CURRENT AND
PROJECTED U.S. USE OF
CFC-11 IN SLABSTOCK FOAM
(In mt)
Year
1985
1990
1995
2000
Slabstock
Foam
459,400
545,600
623,400
712,200
CFC-11
Used
11,500
13,700
15,700
17,900
chloride or possibly to the alternative blowing method is possible. As a
lower bound, we postulate that CFC-11 could lose one-half its share by
2000. Conversely, if the costs of using methylene chloride increase
substantially because of regulation, its share could be taken by CFC-11,
potentially doubling CFC-11 use. Combining these ranges of uncer-
tainty produces a range of uncertainty in 2000 of from 0.48 to 2.07
times the projected base level.
Estimated reporting country use of CFC-11 in slabstock foam is
reported in Table 5.3.2 We estimate non-U.S. use by subtracting
estimated U.S. use derived above. Assuming that non-U.S. use of
CFC-11 in slabstock foam will also grow with the GNP of the respec-
tive countries, we project a future base level as shown in Table 5.4.
The range of uncertainty is similar to that for the United States, albeit
somewhat smaller. Uncertainty about GNP growth implies that total
use could be 20 percent smaller, or 25 percent larger, than the pro-
jected baseline. We believe that total slabstock foam markets, condi-
tional on GNP, could reasonably range between 80 and 125 percent of
the projected level. Finally, since methylene chloride use is smaller
overseas than in the United States and concerns about its possible car-
cinogenicity are perhaps greater, the share of foam blown by CFC-11 is
likely to remain steadier than in the United States: We use an interval
from 0.83 to 1.20. The resulting range for other reporting country
CFC-11 use in slabstock foam falls between 0.69 and 1.44 times the
base projection.
2CMA details reporting country use of CFC-11 in all flexible foams. We estimate use
in molded foam by multiplying world vehicle production by the estimated quantity of
CFC-11 used per vehicle and attribute the remainder to slabstock production.
-------
34
Table 5.3
ESTIMATED HISTORICAL REPORTING COUNTRY
USE OF CFC-11 IN SLABSTOCK FOAM
(In mt)
Year
1976
1977
1978
1979
1980
1981
1982
1983
United
States
11,400
12,900
13,600
13,300
11,700
11,000
9,700
10,600
Other
Reporting
Countries
20,500
26,000
25,500
22,800
26,900
26,500
22,900
31,300
Total
Reporting
Countries
31,900
28,900
39,100
36,100
38,600
37,500
32,600
41,600
Table 5.4
ESTIMATED CURRENT AND PROJECTED
REPORTING COUNTRY USE OF CFC-11
IN SLABSTOCK FOAM
(In mt)
Year
1985
1990
1995
2000
United
States
11,500
13,700
15,700
17,900
Other
Reporting
Countries
33,500
39,800
47,300
56,100
Total
Reporting
Countries
45,000
53,500
63,000
74,000
MOLDED FOAM
Molded foam is used almost entirely for seats and seatbacks in auto-
motive vehicles and its production is intimately related to total vehicle
production. Table 5.5 reports estimated historical U.S. use of CFC-11
in molded foams, molded foam production, and total automobile pro-
duction. Manufacturers report that average use of CFC-11 per unit of
foam has been declining as they have changed product formulations,
substituted methylene chloride, adopted the high resiliency process
-------
35
Table 5.5
ESTIMATED HISTORICAL U.S. PRODUCTION OF AUTOMOTIVE
VEHICLES, MOLDED FOAM AND USE OF CFC-11
Year
1976
1977
1978
1979
1980
1981
1982
1983
1984
CFC-11
(mt)
3,900
4,100
4,300
4,300
3,000
2,800
2,500
3,000
3,200
Molded Foam
(mt)
156,000
164,000
178,000
177,000
131,000
126,000
118,000
145,000
162,000
Automotive
Vehicles
(thousands)
11,500
12,700
12,800
11,400
8,000
8,000
6,900
9,500
10,900
CFC-11 per
Vehicle
(grams)
340
322
333
373
376
347
360
320
297
SOURCE: Vehicles, Automotive News, April 24, 1985.
(which requires no auxiliary blowing agent), and switched to foam den-
sities that can be produced using carbon dioxide alone.3 CFC-11 use
currently averages about 300 grams per vehicle. We combine this fig-
ure with projected growth in vehicle production to project CFC-11 use
in molded foam.
Relying on industry sources, our base projections for vehicle sales in
the United States (total import and domestic sales) grow at 1.9 percent
per year, and sales in foreign markets grow at 2.9 percent. (CFC use is
allocated to the region in which the car is produced though not neces-
sarily sold.) To estimate CFC use outside the United States we assume
that foreign automobiles also average 300 grams of CFC-11 per vehicle.
Table 5.6 summarizes base projections for reporting country vehicle
production and CFC-11 use.
Note that although the base projection for vehicle production grows
at 2.9 and 1.9 percent, the base projection for CFC-11 use grows at
only 0.5 percent outside the United States, and falls 0.4 percent per
year within the United States. This occurs because of the fundamental
uncertainty about future CFC-11 use in molded foam. We believe that
there is little chance that CFC-11 use per vehicle will increase; how-
ever, substantial decreases are possible. Increases are unlikely because
3Molded foams have always used less auxiliary blowing agent than slabstock foam.
Currently, we estimate that molded foams require an average of 20 grams per kilogram.
The average level of CFC-11 use per vehicle is more erratic than CFC-11 use per unit of
foam because of the changing average quantities of foam used per vehicle.
-------
36
Table 5.6
ESTIMATED CURRENT AND PROJECTED REPORTING
COUNTRY AUTOMOTIVE VEHICLE PRODUCTION
AND USE OF CFC-11 IN MOLDED FOAM
United States
Other Reporting
Countries
Total Reporting
Countries
Vehicles CFC-11 Vehicles CFC-11 Vehicles CFC-11
Year (thousands) Used (mt) (thousands) Used (mt) (thousands) Used (mt)
1985
1990
1995
2000
11,100
12,200
13,400
14,700
3,330
3,260
3,190
3,120
29,000
33,400
38,300
44,100
8,700
8,910
9,130
9,360
40,100
45,600
51,800
58,800
12,030
12,170
12,320
12,480
methylene chloride is used much less frequently for molded than for
slabstock foam, so that substitution to CFC-11 would not be as large if
methylene chloride were to be regulated. However, according to indus-
try sources it would be possible to reduce CFC-11 use drastically by
switching to other, unspecified, processes for manufacturing automobile
seats. Because of the possibility of substituting to methylene chloride,
high resiliency or water-blown foam, CFC-11 use per vehicle could
decline by half by 2000. Consequently we project a baseline average
CFC-11 use per vehicle in 2000 to be only 212 grams, with a range of
uncertainty falling between 150 and the current level of 300 grams.
Conditional on GNP, molded foam production could vary by a factor
of about 10 percent, depending on future trends in automobile sizes in
the United States and abroad. Combining these ranges with the range
for GNP growth we obtain an uncertainty range for CFC-11 use in
molded foams in the United States and in other reporting countries of
65 to 153 percent of the base projections.
Figures 5.1 and 5.2 summarize historical and projected use of CFC-
11 in flexible foams in the United States and reporting countries.
Slabstock foam accounts for the majority of reporting country use.
-------
40
en
"O
ca
13
O
CD
03
C
35
30
25
20
15
10
Molded
I I I
I I I
I I I
I I I
1976
1980
1984
1988
Year
1992
1996
2000
Fig. 5.1—Estimated historical and projected U.S. use of CFC-11 in flexible foam
-------
03
to
o
0)
CO
75
c
120
110
100
90
80
70
60
50
40
30
20
10
CO
00
NOTE: Historical molded foam use not available
i i
1976
1980
Molded
i i
1984
1988
Year
1992
1996
2000
Fig. 5.2—Estimated historical and projected reporting country use of CFC-11 in flexible foam
-------
VI. REFRIGERATION AND AIR CONDITIONING
SYSTEMS
CFC-ll and CFC-12 are used as refrigerants in a variety of systems.1
The largest quantities are used in home refrigerators and freezers,
retail store refrigeration systems, chillers (used for cooling buildings)
and mobile (automotive) air conditioners. Except for mobile air condi-
tioners, all of these markets are well developed in the United States
and Western Europe and little growth is expected there. Growth in the
developing countries may be stronger.
The refrigeration cycle consists of compression of the refrigerant fol-
lowed by cooling until it condenses to liquid form, and expansion to
allow evaporation of the liquid. During evaporation the refrigerant
absorbs heat from the refrigeration unit and then returns to the
compressor to repeat the cycle. Since the refrigerant remains in a
sealed system throughout the cycle, the CFC is not emitted until the
unit is disposed of (except for quantities that leak from the system or
are emitted during servicing).
MOBILE AIR CONDITIONING
Air conditioning in automobiles and other vehicles began as a luxury
item in the 1940s and has become commonplace in the United States
but remains a rarity in other countries. Mobile air conditioners use
CFC-12 exclusively. A CFC-22 system was used in the past but has
disappeared, presumably because of the higher cost and weight that
result from the higher pressures required. As with other refrigeration
systems, emissions are generally low until disposal, although significant
emissions occur at servicing and when systems are damaged in automo-
bile accidents.
Use of CFC-12 in mobile air conditioners depends on three factors:
the number of new vehicles, the fraction that are air conditioned, and
the average amount of refrigerant used per system. The fraction of
new cars and trucks sold in the United States with air conditioning has
increased rapidly, from about 6 percent in 1960 to about 60 percent in
the 1970s. Since then that number has remained stable at the current
!When describing refrigerants the CFC prefix is often replaced by the letter R (for
refrigerant) or F (for Freon, a DuPont trademark). Thus CFC-12 is frequently called
R-12 or F-12.
39
-------
40
level of 68 percent.2 The average refrigerant charge in domestic auto-
mobiles declined from about 1.7 kilograms (3.8 pounds) to 1.3 kilo-
grams (2.75 pounds) between 1974 and 1985 as systems were
redesigned to reduce leakage and weight and smaller automobiles
became more common. Further declines are not expected unless vehi-
cles become smaller. Charges in imported automobiles are smaller,
perhaps 0.9 kilograms (2.0 pounds).
Table 6.1 presents historical and projected baseline United States
vehicle sales (total domestic and imported) and estimated CFC-12 used
in mobile air conditioning. Note the decrease in CFC-12 use for 1978
through 1982 because of the decline in initial charge. Annual use
includes the initial charge for new units, manufacturing and installa-
tion losses (about 10 percent of initial charge), replacement of routine
leakage (30 percent of the initial charge is assumed to be lost every 5
years on average), other servicing losses (10 percent of stocks), and
replacement of accident-caused losses (2.5 percent of stocks). The
model for estimating use is described in Palmer et al. (1980). The base
projections assume that United States vehicle sales will grow at 1.9
percent annually through 2000, as industry sources suggest, and that
both the share of new vehicles that are air conditioned and the average
charges will remain constant.
Mobile air conditioning is much less common outside the United
States. We estimate that about 20 percent of non-U.S. vehicles are air
conditioned (65 percent in Japan, 34 percent in Canada, one-half per-
cent in the EEC, and negligible numbers elsewhere).3 Assuming that
average charges and losses are the same as for imports to the United
States (0.9 kilograms average charge), annual use in the other reporting
countries is estimated as 23,000 mt.
Relying on industry sources we project the number of new foreign
vehicles sold to grow at 2.9 percent annually. We further assume that
the share of new vehicles that are air conditioned will grow to 30 per-
cent by 2000, a 50 percent increase from the estimated current level.
Combining these factors leads to a baseline projection of 5.7 percent
annual growth as reflected in Table 6.2.
Uncertainty about future domestic use of CFC-12 in mobile air con-
ditioning is based on uncertainty about total vehicle sales, the share
that are air conditioned, and average charge. We suggest that the
2This is the percentage of all new automotive vehicles, including both automobiles
and trucks, that are air conditioned. The fraction of automobiles is higher, about 82 per-
cent of those produced domestically.
3The Japanese share is reported in Nihon Reito Kucho Kogyo Kai (1985), the EEC
share is from confidential industry sources, and the Canadian share is assumed to be
one-half the United States share because of the colder climate there.
-------
41
Table 6.1
ESTIMATED HISTORICAL AND PROJECTED
U.S. VEHICLE SALES AND USE OF CFC-12
IN MOBILE AIR CONDITIONING
Year
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1990
1995
2000
Vehicles Sold
(millions)
13.28
14.84
15.43
14.13
11.46
10.80
10.54
12.31
14.48
14.76
16.21
17.81
19.57
CFC-12 Use
(mt)
44,890
50,780
52,800
50,910
47,570
47,860
47,670
48,650
50,350
50,400
54,800
59,900
65,400
SOURCE: Historical vehicle sales, Automotive
News, April 24, 1985.
Table 6.2
ESTIMATED CURRENT AND PROJECTED
REPORTING COUNTRY USE OF CFC-12
IN MOBILE AIR CONDITIONING
(In mt)
Year
1985
1990
1995
2000
United
States
50,400
54,800
59,900
65,400
Other
Reporting
Countries
23,000
30,300
40,000
52,800
Total
Reporting
Countries
73,400
85,100
99,900
118,200
uncertainty about vehicle sales is equivalent to that about general
economic growth. The share of new vehicles with air conditioners has
remained at about two-thirds over the last decade, despite substantial
-------
42
changes in the size of automobiles. We suggest a range of uncertainty
of plus or minus 10 percent of the current share. Finally, further
design changes affecting the average charge are not expected, but if the
average size of vehicles changes average charge could vary by perhaps
30 percent. Combining these uncertainties yields an overall range of
uncertainty of 0.70 to 1.43 times the base projection.
Outside the United States, the uncertainty about the share of new
vehicles with air conditioning is greater. Accordingly, we suggest a
range of use conditional on GNP of 0.8 to 1.25 times the base projec-
tion, or an overall range of 0.67 to 1.50 times the base projection.
RETAIL STORE REFRIGERATION
Retail food store refrigeration systems are used to refrigerate the
food and beverages in display cases and to store meat, produce, dairy
products, frozen food, and ice cream in walk-in coolers. Systems are
designed differently depending on whether they are for medium tem-
peratures (storing dairy products) or low temperatures (storing frozen
foods).
System design and choice of refrigerant are influenced by energy
use, since refrigeration equipment accounts for as much as half of a
typical store's energy use. CFC-12 is usually used for medium tem-
perature systems, whereas most low temperature systems use CFC-502,
a blend of nearly equal parts of CFC-22 and CFC-115. Some stores use
CFC-502 for medium temperature systems as well, because of the con-
venience of handling only one refrigerant. However, CFC-502 is sub-
stantially more expensive than CFC-12.
CFC-22 was once employed in both low and medium temperature
systems. Currently it is used only rarely, since it generates excessive
heat in the compressor.
The average refrigerant charge increased during the mid 1970s as a
number of energy saving features were adopted that required additional
refrigerant. This trend will probably continue to some extent until
about 1990. Eventually, the use of low pressure flooding systems in
large supermarkets is expected to be implemented extensively. One
industry source estimates that this will lead to a 10 percent reduction
in average system charge. Another trend that may reduce refrigerant
use is a decline in leakage and servicing emissions as older equipment
is replaced by newer designs.
Table 6.3 presents historical United States data on the number of
retail food stores and their use of CFC-12, and projections. CFC-12
use is estimated using the model described in Palmer et al. (1980),
-------
43
which simulates refrigerant use in four sizes of stores. Annual use
includes refrigerant used for charging new units (between about 60 and
1,100 kilograms, depending on size of store), in testing new units dur-
ing manufacture (about 2 percent of initial charge), and in replacing
losses that occur during installation (5 to 8 percent depending on store
size), leakage, and servicing (about 10 percent of refrigerant in use).
The values in Table 6.3 show a decline in the total number of stores,
but a slight increase in refrigerant use over the period. This reflects a
trend toward larger supermarkets and away from small grocery stores
that is expected to continue in the future. The effect of this trend is
partially offset by expected further substitution to CFC-502 in medium
temperature systems and lower leakage and servicing losses as newer
units with better integrity replace old units.
The base projections for reporting country CFC-12 use are given in
Table 6.4. Industry sources indicate that foreign systems are similar to
those in the United States and that use in the other reporting coun-
tries is currently about equal to United States use. Future growth is
expected to be at the same rate as GNP, 3.5 percent annually, some-
what more rapid than in the United States
Uncertainty about CFC-12 use in retail food stores can be divided
into uncertainty about the number of such stores and about the aver-
age charges and types of refrigerants used. Uncertainty about the
number of stores, conditional on the level of general economic activity,
is small, plus or minus 10 percent. The average charge is highly sensi-
tive to shifts between refrigerants. In recent years there has been some
Table 6.3
ESTIMATED HISTORICAL AND
PROJECTED U.S. USE OF
CFC-12 IN RETAIL FOOD
REFRIGERATION
Year
1976
1980
1985
1990
1995
2000
Number
of Stores
183,700
174,570
164,950
155,330
145,710
136,090
CFC-12
(mt)
4,820
4,830
4,850
4,620
4,910
5,200
-------
44
Table 6.4
ESTIMATED CURRENT AND PROJECTED
REPORTING COUNTRY USE OF CFC-12
IN RETAIL FOOD REFRIGERATION
(In mt)
Year
1985
1990
1995
2000
United
States
4,850
4,620
4,910
5,200
Other
Reporting
Countries
4,850
4,620
5,130
5,670
Total
Reporting
Countries
9,700
9,240
10,040
10,870
movement toward use of CFC-502 in medium temperature systems
because of the convenience of handling only the one refrigerant. CFC-
502 could completely replace CFC-12 in new medium temperature
installations by 2000. If it did CFC-12 use would be reduced by as
much as 75 percent by 2000.4 Alternatively, if the average refrigerant
charge per store continues to increase after 1990, use of CFC-12 in
2000 could be higher by perhaps 20 percent. Combining these uncer-
tainties we suggest a range for CFC-12 use of 0.25 to 1.33 times the
base projections in 2000.
HOME REFRIGERATORS AND FREEZERS
In the early part of this century refrigerants such as methyl chloride,
ammonia, and sulfur dioxide were employed in home refrigeration
equipment. Since these chemicals are toxic, and some are also flam-
mable or explosive, they are not considered safe enough for home use.
CFC-12 has none of these drawbacks and has been used in all home
refrigerating equipment in the United States since at least 1946.
Home refrigerators currently use either a rotary or reciprocating
compressor. The reciprocating-compressor machines require only one-
third to one-half as much refrigerant. Historically, the average charge
in a domestic refrigerator was 280 grams (10 ounces) and the average
4Leakage and servicing use represent about half of refrigerant use. Older units using
CFC-12 would still require CFC-12 for servicing. If half the medium temperature units
in service used CFC-502 by then, CFC-12 use would still account for one-fourth the pro-
jected use.
-------
45
charge in a freezer was 430 grams (15 ounces). Today, refrigerators use
an average of about 210 grams (7.5 ounces) of refrigerant, whereas
freezers require about 280 grams (10 ounces).5 Although refrigerators
outside the United States are typically smaller, we assume that initial
charges are the same as in current U.S. practice, about three-quarters
the historic U.S. level. We assume that home freezer use outside the
United States is negligible.
CFC-12 is used in home appliances for charging new units, in testing
and rework at manufacturing, and in replacing leakage and servicing
losses. The models used to calculate historical and future use are
described in Palmer et al. (1980).
Table 6.5 reports historical and projected baseline U.S. sales of
home appliances and estimated use of CFC-12. CFC-12 use declined
between 1976 and 1984 because of the increased use of reciprocating
compressors in refrigerators, which require a smaller initial charge, and
because the growth in freezers declined. Since the domestic home
refrigerator market is well saturated, we project future growth slightly
below the expected growth of GNP, 3 percent to 1990 and 2.5 percent
to 2000. Freezer sales were strong in the early 1970s but have since
declined. We assume only 1 percent growth to 2000.
Table 6.5
ESTIMATED HISTORICAL AND PROJECTED
U.S. HOME REFRIGERATOR AND FREEZER
SALES AND USE OF CFC-12
Year
1976
1980
1984
1985
1990
1995
2000
Number of
Refrigerators
(thousands)
5,090
5,570
6,200
6,390
7,410
8,380
9,480
Number of
Freezers
(thousands)
1,740
1,670
1,340
1,360
1,430
1,500
1,570
CFC-12
(mt)
2,920
2,370
2,400
2,440
2,660
2,900
3,190
SOURCES: Palmer et al. (1980), Mooz et al.
(1982).
Home refrigerators typically incorporate three to six times more CFC-11 blowing
agent in the insulating foam in their walls than CFC-12 refrigerant in their cooling sys-
tems.
-------
46
Both reciprocating and rotary compressors are competitive in the
home appliance market. However, if their market shares were to shift
significantly it could substantially affect CFC-12 use. At present,
about 25 percent of domestic refrigerators and 30 percent of freezers
contain rotary compressors. If the share of appliances using rotary
compressors were to halve or double, average initial charges could fall
25 percent or increase by one-third by 2000. Combined with the uncer-
tainty over the total number of home refrigerators and freezers pro-
duced (equivalent to the uncertainty about GNP) the range of uncer-
tainty is from 0.7 to 1.43 times the projected base for CFC-12 use in
2000.
Table 6.6 reports estimated home refrigerator production by world
region. As a baseline projection we assume that West European use
will halt its historical decline and remain constant at 11,000 mt annu-
ally to 2000. For the other reporting countries we set baseline pro-
jected growth at the same rate as from 1972 to 1981, 6.2 percent per
year. 6
Table 6.7 reports our base projection for reporting country CFC-12
use in home appliances. Uncertainty about use outside the United
States is greater, because of the greater uncertainty about the average
charge and how it may change. We suggest a range of 0.67 to 1.50
Table 6.6
ESTIMATED HISTORICAL WORLD PRODUCTION
OF HOME REFRIGERATORS
(In thousands)
Region
Western Europe
North America (excl. U.S.)
South America
Asia (excl. USSR)
Africa
Oceana
Eastern Europe (excl. USSR)
USSR
Total
1972
12,110
830
1,330
4,550
330
450
2,510
5,030
27,100
Average growth
1981 rate (%)
10,990
1,350
2,800
7,710
510
460
3,340
5,930
33,100
-1.1
5.6
8.6
6.0
5.0
0.0
3.2
1.8
2.2
SOURCE: United Nations Statistical Office (1981).
6Recall that we do not project use in the communist countries in this section; the data
on communist countries in Table 6.6 is included for reference.
-------
47
times the base projection in 2000 to account for the uncertainty about
average charge. Combined with uncertainty about GNP growth this
produces a range of uncertainty of 0.63 to 1.59 times projected use.
CHILLERS
Chillers are air conditioning systems used in large commercial and
industrial buildings. The cooling system consists of a central unit that
chills a secondary refrigerant, typically water or brine, which is circu-
lated to cooling coils throughout the building. Chillers may use either
centrifugal or reciprocating compressors. The centrifugal units use pri-
marily CFC-11 or CFC-12 and have capacities ranging from 75 to
10,000 tons.7 The reciprocating units use CFC-22 and CFC-12 and are
of smaller capacity, usually no more than 150 tons.
Estimates of historical and projected baseline CFC-11 and CFC-12
use are developed using a model described in Palmer et al. (1980). Use
includes the CFC used for charging new units, for testing at manufac-
ture (about 5 percent of initial charge, falling over time), for replacing
losses during shipping and installation (4 percent of initial charge), and
for replacing leakage and servicing losses (about 7.5 and 10 percent of
banked refrigerant, respectively; estimated servicing losses decline over
time).
Table 6.7
ESTIMATED CURRENT AND PROJECTED
REPORTING COUNTRY USE OF CFC-12 IN
HOME REFRIGERATORS AND FREEZERS
(In mt)
Year
1985
1990
1995
2000
United
States
2,440
2,660
2,900
3,190
Other
Reporting
Countries
7,790
10,520
14,220
19,200
Total
Reporting
Countries
10,230
13,180
17,120
22,390
'Refrigerating system capacities are traditionally rated in tons, where one ton is the
amount of heat required to melt one ton of ice in 24 hours or 200 Btu per minute. A
commercial or industrial building generally requires about one ton of cooling capacity for
every 300 square feet of floor area.
-------
48
Centrifugal Chillers
The choice of refrigerant in a centrifugal chiller is also a matter of
size. CFC-11 is used in about 80 percent of the units, but these are
primarily small units (less than 500 tons), requiring an estimated 350
kilogram average charge. CFC-12 is used in about 10 percent of the
units, with an average charge estimated as 950 kilograms. The largest
units (over 1,000 tons) are usually charged with CFC-22, CFC-114, or
CFC-500—a blend containing 74 percent CFC-12. Average charges are
declining slowly as manufacturers make evolutionary advances in
chiller design and reduce leakage.
Table 6.8 presents historical estimates of domestically installed cen-
trifugal chillers and attendant CFC-11 and CFC-12 use and our base
projection for future use. The number of chillers grew steadily between
1976 and 1981 and then declined slightly, perhaps because of the reces-
sion. CFC use over the period increased by only about 10 percent
because of a decrease in servicing losses from an estimated 16 percent
of stocks in 1976 to about 10 percent in 1985. Note the slower growth
in CFC use after 1985 because of projected further decreases in servic-
ing losses, to 5 percent of stocks by 2000.
Table 6.8
ESTIMATED HISTORICAL AND PROJECTED U.S.
CENTRIFUGAL CHILLER INSTALLATIONS
AND USE OF CFC-11 AND CFC-12
Year
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1990
1995
2000
Centrifugal
Chillers
1,930
1,910
2,280
2,350
3,090
3,580
3,210
3,470
3,420
3,580
4,490
5,260
5,750
SOURCES: Chillers,
CFC-11
(mt)
3,910
3,900
4,010
4,020
4,260
4,440
4,330
4,400
4,360
4,380
4,840
4,950
5,400
1976-1980,
CFC-12
(mt)
1,460
1,430
1,480
1,480
1,590
1,670
1,620
1,650
1,640
1,650
1,800
1,830
2,000
Mooz et al.
(1982); 1981-1984, industry sources.
-------
49
One trend that may reduce future use of CFC-11 in centrifugal chill-
ers is the increasing use of one or more reciprocating chillers, most of
which use CFC-22, instead of a single centrifugal unit. This trend is
expected to continue as larger reciprocating units are manufactured.
The use of multiple reciprocating chillers in place of a single centrifu-
gal unit provides greater system reliability and allows continued cooling
while one unit is being serviced. If this trend were to continue, use of
CFC-11 might be reduced by one-third by the year 2000.8
We estimate current reporting country use relying on industry
sources who suggest that the United States currently accounts for
about 45 percent of CFC-11 and CFC-12 use in centrifugal chillers.
Table 6.9 reports projected reporting country use. We assume that
chiller sales for new construction grow at the rate of GNP both in the
United States and abroad.
Uncertainty concerning future use of CFCs in centrifugal chillers
centers on possible substitution to reciprocating chillers. We suggest
that uncertainty about the total demand for cooling systems is compar-
able to uncertainty over GNP growth, whereas the quantities of CFC-
11 and CFC-12 used in centrifugal chillers could decline by one-third
by 2000 if reciprocating chillers take a significant share of the centrifu-
gal market. These uncertainties combine to a range of 63 to 125 per-
cent of the base projections.
Table 6.9
ESTIMATED CURRENT AND PROJECTED REPORTING
COUNTRY USE OF CFC-11 AND CFC-12
IN CENTRIFUGAL CHILLERS
(In mt)
Other Reporting Total Reporting
Year
1985
1990
1995
2000
United
CFC-11
4,380
4,840
4,950
5,400
States
CFC-12
1,650
1,800
1,830
2,000
Countries
CFC-11
5,520
6,090
6,510
7,420
CFC-12
2,080
2,270
2,410
2,750
Countries
CFC-11
9,900
10,930
11,460
. 12,820
CFC-12
3,730
4,070
4,240
4,750
8If reciprocating chillers take half the market we project for centrifugal units, the
stock of centrifugal chillers would be 20 percent smaller than projected. Since replace-
ment of leakage and servicing losses account for 68 percent of CFC-11 use in 2000, use
would be only about two-thirds the projected level.
-------
50
Reciprocating Chillers
Smaller cooling systems, of less than about 150 tons capacity, typi-
cally employ reciprocating chillers. Although the systems were origin-
ally designed to use CFC-12, CFC-22 designs have been replacing the
older units since the 1960s; today only about 2 percent of new units use
CFC-12. The CFC-22 systems have about a 60 percent greater capacity
than CFC-12 systems for the same physical size.
Table 6.10 presents estimated historic U.S. sales of reciprocating
chillers and associated CFC-12 use and the base projection for future
use. Note that in spite of the growth in the number of chillers, CFC-12
use declined precipitously as an increasing share of chillers used CFC-
22. This decline is projected to continue as older units that use CFC-
12 are replaced by new CFC-22 using units.
Table 6.11 reports the base projection for reporting country CFC-12
use in reciprocating chillers. Estimated current use is based on indus-
try sources indicating that reporting country use of CFC-12 in recipro-
cating chillers is currently 3.6 times as large as U.S. use. We believe
that growth in new reciprocating chillers outside the United States will
be somewhat more rapid, 5 percent annually, because most other coun-
tries have concentrated on reciprocating instead of centrifugal chillers
because of the lower production capital costs.
Uncertainty over total demand for chillers is comparable to uncer-
tainty over GNP growth. Moreover, the level of CFC-12 use in chillers
Table 6.10
ESTIMATED HISTORICAL AND
PROJECTED U.S. RECIPROCATING
CHILLER INSTALLATIONS AND
USE OF CFC-12
Year
1976
1977
1978
1979
1980
1985
1990
1995
2000
Reciprocating
Chillers
5,020
6,970
6,300
6,850
8,020
9,300
10,350
11,100
12,700
CFC-12
(mt)
890
830
750
670
600
280
130
90
90
SOURCE: 1976-1980, Mooz et al.
(1982).
-------
51
Table 6.11
ESTIMATED CURRENT AND PROJECTED
WORLD USE OF CFC-12 IN
RECIPROCATING CHILLERS
(In mt)
Year
1985
1990
1995
2000
United
States
280
130
90
90
Other
Reporting
Countries
1,000
510
400
450
Total
Reporting
Countries
1,280
640
490
540
could vary widely. If all new chillers used CFC-22, use of CFC-12
could decline to about half the base projection by 2000. Alternatively,
if centrifugal chillers continue to be replaced by reciprocating units and
significant substitution of CFC-22 does not occur, CFC-12 use could
conceivably double. Combining these uncertainties produces a range of
possible use in 2000 of 48 to 207 percent of the base projections.
Figures. 6.1 and 6.2 summarize the projected baseline use of CFC-11
and CFC-12 in refrigeration applications. Almost all of the use we
have analyzed is of CFC-12, and the United States accounts for the
majority of analyzed use because of widespread mobile air conditioning.
-------
100
90
80
70
S 60
T3
CO
o 50
0>
(A
| 40
30
20 -
10 -
J L
1976
1980
CFC-11
J 1 1 I I I I
1984
1988
Year
J I I
1992
1996
Fig. 6.1—Estimated historical and projected U.S. use of CFC-11 and CFC-12
in refrigeration and air conditioning
Ol
10
2000
-------
tf>
~a
c
(0
en
o
0)
tfi
"55
c
c
20 -
1985
1990
1995
2000
Year
Fig. 6.2—Estimated current and projected reporting country use of CFC-11 and CFC-12
in refrigeration and air conditioning
Ol
CO
-------
VII. MISCELLANEOUS USES OF CFC-11 AND
CFC-12
There are a variety of other applications that use small amounts of
CFC-11 and CFC-12. Although the amount of CFC used in each speci-
alty application alone is small, taken together such uses represent a
significant fraction of CFC use. In what follows, we estimate the
present and future use of CFC-11 and CFC-12 in miscellaneous appli-
cations. We do not attempt to construct ranges of uncertainty for
these minor uses and present only point estimates for future years.
STERILANTS
CFC-12 is used as a diluent for ethylene oxide in hospital and indus-
trial sterilants. In earlier work (Palmer et al., 1980; Mooz et al., 1982),
Rand estimated U.S. use of CFC-12 in sterilants at 5,900 mt in 1976
and 6,800 mt in 1980. During that four year period the growth aver-
aged 3.6 percent annually. Assuming the same rate of growth for the
next five years would place 1985 use at about 8,000 mt.
Industry sources indicate that use of the ethylene oxide/CFC-12
blend will not increase in the future. Indeed, ethylene oxide, an animal
carcinogen, may eventually be banned in sterilant applications. Accor-
dingly, we assume no future growth. We arbitrarily assume that CFC-
12 use in sterilants outside the United States accounts for an addi-
tional 8,000 mt and that this use will not grow.
LIQUID FOOD FREEZING
CFC-12 is used in a liquid food freezing process developed by
DuPont. 1976 U.S. use was estimated at about 2,720 mt, with large
future growth anticipated (Palmer et al., 1980). By 1980, the expected
growth had not occurred and future growth was estimated at much
lower levels (Mooz et al., 1982).
Using this information we estimate an average annual growth of 1
percent since 1976, placing 1985 use at 2,970 mt. Use in the other
reporting countries is also small: We estimate that it is about equal to
U.S. use.
%
Table 7.1 presents estimated CFC-12 use in liquid food freezing for
the United States and other reporting countries. The projections were
54
-------
55
Table 7.1
ESTIMATED CURRENT AND PROJECTED USE OF
CFC-12 IN LIQUID FOOD FREEZING
(In mt)
Year
1985
1990
1995
2000
United
States
2,970
3,530
4,030
4,600
Other
Reporting
Countries
2,970
3,530
4,190
4,980
Total
Reporting
Countries
5,940
7,060
8,220
9,580
derived assuming future growth at the projected regional GNP growth
rates. Industry sources indicate that new LFF users are few and that
CFC-12 use for this purpose may actually decline in the future.
OTHER MISCELLANEOUS APPLICATIONS
A variety of other applications use very small amounts of CFC-11
and CFC-12. They include CFC-12 use in fire and other warning
devices, boat horns, dehumidifiers, pressurized blowers and drain
cleaners, and trucking refrigeration, and CFC-11 use in coal cleaning.
Total U.S. use in these applications was estimated at 450 mt of CFC-
11 and 1,430 mt of CFC-12 in 1976 (Palmer et al., 1980). In Table 7.2
we provide estimates of current and future U.S. use based on growth at
the GNP rate for all applications except coal cleaning. Industry
sources indicate that this use has not grown and is unlikely to do so in
the future. The table also presents estimated current and projected use
of CFC-12 in the other reporting countries assuming that 1985 levels
mirror those in the United States and that growth will occur at the
same rate as the GNP. We do not include CFC-11 since we are not
aware of its use for coal cleaning outside the United States.
-------
56
Table 7.2
ESTIMATED CURRENT AND PROJECTED USE OF CFC-11 AND
CFC-12 IN OTHER MISCELLANEOUS APPLICATIONS
(In mt)
Other Reporting Total Reporting
Year
1985
1990
1995
2000
United
CFC-11
450
450
450
450
States
CFC-12
1,700
2,020
2,301
2,640
Countries
CFC-11 CFC-12
— 1,700
— 2,020
- 2,400
— 2,850
Countries
CFC-11
450
450
450
450
CFC-12
3,400
4,040
4,710
5,490
-------
VIII. SOLVENTS
CFC-113 and methyl chloroform are used primarily as solvents.
Carbon tetrachloride was formerly used as a solvent but, at least in the
industrialized countries, this use has been curtailed. It is also used as a
grain fumigant, but that use has been discontinued in the United
States. The chemical's major use is as an intermediate in the produc-
tion of CFC-11 and CFC-12.
The use of solvents as "degreasers" is an important cleaning step in
many manufacturing processes. Projections of future solvent use are
especially uncertain because of the number of competing solvents and
the possibilities for substitution among them as one or more are sub-
jected to increasing regulation. Neither CFC-113 nor methyl chloro-
form has a dominant share of the overall solvent market, although each
has a strong position in certain submarkets. CFC-113 is used primarily
in the electronics industry, and methyl chloroform is used in a variety
of applications from electronics to shipbuilding.
Most solvent uses can be described as either "cold cleaning," where
the part to be cleaned is dipped into a tank of solvent, or "vapor
degreasing," where the solvent is heated and the item to be cleaned is
suspended in the vapor above the tank. In both types of cleaning the
solvent displaces contaminants and then evaporates. Another use is
drying, where the solvent is used to displace water. Solvent emissions
are prompt, although equipment to recover and recycle the solvent can
be used.
Use of methyl chloroform and CFC-113 may be strongly affected by
government regulations. One factor that may discourage the use of
these and other solvents is the U.S. ban on land disposal of waste
chlorinated solvent, which becomes effective in November 1986. Some
users may adopt more conservative practices or alternative cleaning
processes, or substitute other solvents that have not been banned from
land disposal to avoid much higher disposal costs. Alternatively, some
of the competing solvents, such as trichloroethylene, perchloroethylene
and methylene chloride, may be regulated more stringently because of
concern over health risks to workers and the general population.
(These solvents have been shown to cause chronic adverse health
effects in animal studies.) Such restrictions could substantially
increase demand for CFC-113 and methyl chloroform.
57
-------
58
CFC-113
CFC-113 is used largely in the electronics industry, to deflux printed
circuit boards and other metal components and for "critical cleaning"
of plastic or specialty components, such as semiconductors, that are
produced in a contamination-controlled environment. CFC-113 is more
compatible with certain plastics than are other solvents: Without it
many components could not be made of plastic. Small quantities are
also used in chemical processing, as a blowing agent for foam produc-
tion, as a specialty refrigerant, and for dry cleaning.
CFC-113 has several advantages over potential alternatives. First, it
is a mild solvent, compatible with virtually all materials. Indeed, it is
gentle enough to dry clean delicate suede and leather garments.
Second, its TLV1 in the workplace is 1,000 parts per million, the
highest value assigned. This implies that costly methods of air circula-
tion and dilution can be avoided. Third, CFC-113 is very stable, one of
the characteristics that makes it a potential ozone depleter. Thus it
can be used without stabilizing additives.
In Table 8.1 we present estimates of historical U.S. CFC-113 pro-
duction and projected baseline future use. Because the chemical is pro-
duced by only two domestic producers, annual production data are not
reported by the International Trade Commission. We estimated his-
torical use as follows: The 1976 and 1977 values are from Palmer et al.
(1980). The 1978 estimate was derived from a Chemical Marketing
Reporter Chemical Profile on fluorocarbons indicating that solvent use
amounted to 11 percent of total CFC use (CMR, 1978). Since virtually
all solvent use is CFC-113, we estimate 1978 use as 11 percent of 1978
CFC production. Total CFC production is estimated from the ITC
reports for CFC-11, CFC-12, and CFC-22, and a Rand estimate for
CFC-114.2
For 1979, we based our estimate on a Chemical Marketing Reporter
Chemical Profile on perchloroethylene, the precursor chemical used to
produce CFC-113 and CFC-114 (CMR, 1979). The profile indicated
that 13 percent of the chemical was used in CFC production. After
subtracting the estimated amount used to produce CFC-114 the
amount of CFC-113 produced can be estimated using the appropriate
stochiometric equations (see Wolf, 1980).
lrThe Threshold Limit Value is the maximum allowable time-weighted average con-
centration to which a worker may be exposed over an eight-hour working day, 40-hour
work week, as determined by the U.S. Occupational Safety and Health Administration
(OSHA). Generally, the more toxic the chemical, the lower the TLV.
Total CFC-113 use exceeds solvent use by the small amounts used as refrigerants
and in plastic foams.
-------
59
Table 8.1
ESTIMATED HISTORICAL
AND PROJECTED U.S.
USE OF CFC-113
(In mt)
Year
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1990
1995
2000
CFC-113
31,300
36,740
39,010
49,900
57,150
57,150
57,150
57,150
68,040
73,200
105,600
134,800
172,000
The 1984 production estimate is from industry sources. The values
for 1980 through 1983 are more uncertain. CFC-113 production during
that time may have been as high as 59,870 mt (EPA, 1981) or as low as
57,150 mt (industry sources). We adopt the latter estimate and assume
that the market was flat for the period because of the recession.
The estimates in Table 8.1 suggest that CFC-113 use more than
doubled between 1976 and 1984. We expect future growth to be high
as well. In the absence of the ban on land disposal of waste chlori-
nated solvents, CFC-113 use in defluxing and critical cleaning (about
50 percent of current use) would be expected to grow at the rate of
other chemicals used in the electronics industry, 13 percent over the
next several years.3 As a base projection, we assume that defluxing and
critical cleaning use grow at only 10 percent between 1984 and 1990 to
account for the land disposal ban. We assume that the other CFC-113
applications grow at a lower rate, 5 percent through 1990. Combining
these markets implies overall growth averaging 7.6 percent annually.
Because industry sources expect the electronics industry to continue to
grow more rapidly than GNP for an extended period, we project base-
line CFC-113 use in all applications to grow at an average of 5 percent
annually between 1990 and 2000.
3See Stinson (1983), CMR (1984), Mining Journal (1985), and Manufacturing Chemist
and Aerosol News (1985).
-------
60
Industry sources suggest that 1984 CFC-113 use in the EEC was less
than 45,000 mt; we assume 40,000 mt. Japanese production of all
CFCs is about 150,000 mt4 of which CFC-11 and CFC-12 account for
110,000 mt.5 Almost all of the remaining 40,000 mt should be CFC-113
and CFC-114; we assume that CFC-113 production is 35,000 mt.
Finally, we assume that production in the other reporting countries
outside the EEC and Japan is nominal, about 10,000 mt. This leads to
an estimated 85,000 mt in the CMA reporting countries outside the
United States.
Our base projections for reporting country CFC-113 production are
reported in Table 8.2. Outside the United States, we project baseline
growth of 9.3 percent for the period 1984 through 1990. This rate is
based on one-half the market (defluxing and critical cleaning) growing
at 13 percent, as U.S. use would in the absence of the land disposal
ban, and the other half at 5 percent. After 1990, we assume an average
annual rate of 6 percent growth in the baseline, about 70 percent
higher than the expected GNP growth rate, to account for the fast
growing electronics industry.
United States CFC-113 use may be lower than our projected base
use if the impending ban on land disposal affects CFC-113 use more
than we have estimated. Two of the most likely alternatives to land
disposal are reclamation and incineration. Once the ban goes into
effect there may be insufficient reclamation capacity to handle the
greatly increased demand. Indeed, today only a handful of reclaimers
in the country can properly reclaim CFC-113. The fluorine in CFC-113
Table 8.2
ESTIMATED CURRENT AND PROJECTED
REPORTING COUNTRY USE OF CFC-113
(In mt)
Year
1985
1990
1995
2000
United
States
73,200
105,600
134,800
172,000
Other
Reporting
Countries
85,000
132,600
177,400
237,500
Total
Reporting
Countries
158,200
238,200
312,200
409,500
4Ministry of International Trade and Industry data as supplied by Japan Flon Gas
Association (private communication, 1985).
5Taya (1985).
-------
61
apparently corrodes the refractory material lining incinerators, and
industry sources report that incinerator owners charge dearly for fluori-
nated chemicals or refuse to accept them altogether. Consequently, it
is possible that substitution and increased reclamation might reduce
U.S. CFC-113 production by a third. Alternatively, regulation of tri-
chloroethylene, perchloroethylene. and methylene chloride use could
increase demand for CFC-113 by half. We believe that outside the
United States any regulations are likely to have a smaller effect, since
the existence and stringency of regulations will differ across nations.
However, future growth in solvent use is more uncertain so we also
assume a range of plus one-half or minus one-third.
METHYL CHLOROFORM
Although it is not a particularly strong potential ozone depleter,
methyl chloroform is produced in large quantities. In the United
States, it is used primarily for vapor degreasing and cold cleaning of
electronic and other parts, although small amounts are used in
adhesives, aerosols, and coatings. Methyl chloroform is a general pur-
pose solvent that has certain advantages over its competitors. Its
TLV, although lower than that of CFC-113, is higher than that of
other chlorinated solvents. It is a stronger solvent than CFC-113 and
is consequently not compatible with all materials. However, because it
is stronger it can clean contaminants that CFC-113 cannot remove. A
recently marketed blend of methyl chloroform and alcohol has made
some inroads into the electronics market.
Table 8.3 reports U.S. methyl chloroform production between 1976
and 1984 and projected use. Note that production increased 16 percent
over the period, an average of about 1.9 percent annually.
We estimate 1985 U.S. use of methyl chloroform as 270,000 mt.
This estimate is based on the trend in reported production since 1979
and is derived in Appendix A. Since methyl chloroform is a widely
used general purpose solvent we would expect use to grow at the same
rate as the GNP in the absence of the land disposal ban. Because of
the ban we assume a slightly lower growth, 3 percent between 1986 and
1990 and 2 percent through 2000 (compared with expected GNP
growth of 3.5 and 2.7 percent, respectively). Before the ban's imple-
mentation in 1986 we assume growth at the GNP rate.
Industry sources indicate that 1980 U.S. production represented 59
percent of total world production outside the communist countries. If
we assume this same relationship for 1985, we obtain the values shown
in Table 8.4. We project baseline growth outside the United States at
the same rate as GNP, 3.5 percent annually.
-------
62
Table 8.3
ESTIMATED HISTORICAL
AND PROJECTED U.S.
USE OF METHYL
CHLOROFORM
(In mt)
Year
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1990
1995
2000
SOURCE:
(1976-1984).
Methyl
Chloroform
266,000
288,000
292,000
325,000
314,000
279,000
270,000
266,000
303,000
270,000
314,500
347,300
383,400
U.S. ITC
Projected methyl chloroform use is subject to the same types of
uncertainty as CFC-113 use. The land disposal ban may cause users to
reduce their use of methyl chloroform (through reclamation, for
Table 8.4
ESTIMATED CURRENT AND PROJECTED REPORTING
COUNTRY USE OF METHYL CHLOROFORM
(In mt)
Year
1985
1990
1995
2000
United
States
270,000
314,500
347,300
383,400
Other
Reporting
Countries
187,600
222,800
264,600
314,300
Total
Reporting
Countries
457,600
537,300
611,900
697,700
-------
63
example) or to switch to other solvents or processes. Alternatively,
regulation of alternative chlorinated solvents, including methylene
chloride, trichloroethylene, and perchloroethylene, could substantially
increase domestic use of methyl chloroform. Since methyl chloroform
is used in a wider range of applications than CFC-113, its use is less
sensitive to trends in particular industries and the range of uncertainty
should be smaller. We suggest a range of uncertainty of 80 to 125 per-
cent of the base projection, conditional on GNP. Combined with the
uncertainty about GNP growth, the overall range of uncertainty is 0.73
to 1.37 times the base projection for the United States and for the
other reporting countries.
CARBON TETRACHLORIDE
Carbon tetrachloride is an excellent solvent in many applications.
At one time, it was widely used as a solvent in the United States.
Because of its acute and chronic toxicity, however, it is used only in
small amounts for such purposes today. Although our information on
carbon tetrachloride use in the rest of the world is limited, we suspect
that carbon tetrachloride is commonly employed as a general purpose
solvent in many developing nations.
Carbon tetrachloride's major use in the United States and most of
the rest of the world is as an intermediate in the production of CFC-11
and CFC-12. A second large use in this country has been for grain
fumigation. Domestic producers have voluntarily agreed to halt pro-
duction for this application beginning in 1986, however. Additional
small amounts are used in the pharmaceutical industry.
Table 8.5 shows historical U.S. production of carbon tetrachloride
and the fraction used to produce CFCs. Note the sharp decline in 1982
and 1983 because of the recession. Demand for CFC-11 and CFC-12
was off strongly for that period, and because a high percentage of car-
bon tetrachloride is devoted to CFC manufacture, its total production
declined as well (CMR, 1983).
Table 8.6 presents estimated future carbon tetrachloride production
using the base projections for United States and other reporting coun-
try CFC-11 and CFC-12 production. The estimates are based on the
stochiometric equations for CFC production. These imply that produc-
tion of one kilogram of CFC-11 requires 1.12 kilograms of carbon tetra-
chloride and that one kilogram of CFC-12 requires 1.27 kilograms (see
Wolf, 1980). We estimate that additional losses in the production pro-
cess amount to 2.7 percent of total use in CFC production, whereas
other uses, not including grain fumigation, account for an estimated
-------
64
Table 8.5
ESTIMATED HISTORICAL U.S. USE
OF CARBON TETRACHLORIDE
(In mt)
Year
1976
1977
1978
1979
1980
1981
1982
1983
1984
Carbon
Tetrachloride
389,000
366,000
334,000
324,000
322,000
329,000
266,000
260,000
323,000
% Used To
Produce CFCs
—
—
95
—
—
95
—
83.6a
—
SOURCES: U.S. ITC (1976-1984);
CMR (1978, 1981, 1983).
"The share allocated to CFC produc-
tion decreased in 1983 because a large
fraction of carbon tetrachloride was
exported.
Table 8.6
ESTIMATED CURRENT AND PROJECTED REPORTING
COUNTRY USE OF CARBON TETRACHLORIDE
(In mt)
Year
1985
1990
1995
2000
United
States
280,000
323,200
373,100
430,700
Other
Reporting
Countries
590,000
668,500
757,300
857,700
Total
Reporting
Countries
870,000
991,700
1,130,400
1,288,400
additional 5.4 percent of total use. The resulting U.S. estimate, 277,000
mt, almost exactly matches the 280,000 mt estimate derived from
reported production in recent years (see Appendix A). Since almost all
carbon tetrachloride is used for CFC production the uncertainty about
future production is based on uncertainty about future production of
CFC-11 and CFC-12. Specifically, the range is based on the average of
-------
65
the ranges of total CFC-11 and CFC-12 production and spans the
interval from 0.76 to 1.32 times the projected base level in the United
States and from 0.73 to 1.37 times the base projection for the other
reporting countries.
Historical and projected consumption of CFC-113, methyl chloro-
form, and carbon tetrachloride are summarized in Fig. 8.1 for the
United States and in Fig. 8.2 for the CMA reporting countries. Note
that U.S. use of methyl chloroform about equals that of carbon tetra-
chloride, but carbon tetrachloride use far exceeds estimated methyl
chloroform use in the reporting countries. This difference is due in
part to the relatively low U.S. use of CFC-11, compared with the other
reporting countries, and consequently lower carbon tetrachloride use.
-------
700
0 I i i i I
1976
05
2000
Fig. 8.1—Estimated historical and projected U.S. use of CFC-113,
methyl chloroform, and carbon tetrachloride
-------
CO
CO
O
0)
W
1985
1990
1995
2000
Year
Fig. 8.2—Estimated current and projected reporting country use of CFC-113,
methyl chloroform, and carbon tetrachloride
-------
IX. FIRE EXTINGUISHANTS
Halons 1211 and 1301 were introduced for use as fire extinguishants
in the United States in the early 1970s. Although total production is
presently small relative to the other chemicals we consider, use is
expected to grow rapidly. Moreover, since the Halons contain bromine
they may present a much greater threat to the stratospheric ozone than
do similar chlorinated chemicals.
Halon 1301 is used primarily in total flooding systems that protect
valuable equipment and materials in enclosed spaces. These systems
release a preset quantity of the gas from a pressurized cylinder in the
event of a fire. Although Halon 1301 is relatively expensive it leaves
no residue and does not damage valuable equipment. Moreover, unlike
other extinguishants, it can be released before workers are evacuated
from the room, thereby reducing fire damage. It is used primarily in
'computer rooms, telephone exchanges, pipeline compressor and pump-
ing stations, airliners, some libraries and museums, battle tanks, and
ship engine and boiler rooms. In the late 1970s more than 80 percent
was used in computer rooms, telephone exchanges, and other rooms
containing electronic equipment (DuPont, 1978). Halon 1301 is also
being introduced into the hand-held extinguisher market for home and
commercial use, and small quantities are used as a specialty refriger-
ant.
Halon 1211 is used primarily in hand-held fire extinguishers. It is
also used in U.S. Air Force rapid intervention crash trucks.
Nearly the entire quantity of Halons produced in a year is banked in
total flooding systems or other fire extinguishers. Emissions occur
when the system is activated during a fire and from system testing, fill-
ing and servicing, leakage, and accidental discharges. Losses through
accidental discharge are likely to be more significant for hand-held
extinguishers than for the total flooding systems. Since the Halons are
so expensive it is likely that the amounts stored in total flooding sys-
tems will often be recovered and re-used when the system is disman-
tled, so much of the bank may never be released to the atmosphere.
68
-------
69
HISTORICAL U.S. USE
Table 9.1 presents estimated historical U.S. use of Halon 1301. We
have little information on past use of Halon 1211, but suspect it is
similar. Approximately 95 percent of annual Halon 1301 use is attri-
buted to total flooding systems with the remaining 5 percent going to
hand-held extinguishers, specialty refrigeration, and other uses.
The historical use estimates assume the introduction of 2,600 new
systems each year beginning in 1972 and reaching a total of 13,000 sys-
tems in 1978. Subsequent growth was rapid: an estimated 30 percent
annually during 1978 and 1979, falling to 25 percent in 1980 and 1981,
to 15 percent in 1982 and 1983, and recovering to 20 percent in 1984
with the end of the recession. The estimates are consistent with esti-
mates by industry sources that U.S. use in 1977 was between 1,100 and
1,400 mt and that use in recent years was 3,400 to 4,500 mt.
Although there is considerable uncertainty about the average system
charge, industry sources indicate that it was quite high in the early
years and has since fallen. Accordingly, we adopt an average system
charge of 340 kilograms (750 pounds) for all new systems through 1983,
falling to 318 kilograms in 1984.
Table 9.1
ESTIMATED HISTORICAL
U.S. USE OF
HALON 1301
(In mt)
Year
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
Use
1,200
1,200
1,200
1,200
1,200
1,600
2,100
2,600
3,200
3,700
4,300
4,800
-------
70
The estimated use figures in Table 9.1 include: the Halon 1301 used
to charge new units; use for testing, estimated at 19 percent of the ini-
tial charge; replacement of filling and servicing losses, estimated at 1
percent of the initial charge; replacement of leakage, estimated at 0.1
percent of the bank; replacement of Halon 1301 at system failure,
estimated at 1 percent of the bank; and discharge during fire, estimated
at 1 percent of the bank.
Halon 1301 use for system testing may increase substantially from
its current estimated use of 19 percent of initial charge. Currently,
three-quarters of new systems are tested, three-quarters of these using
CFC-12, but installers are beginning to test with the Halon more fre-
quently for two reasons. First, although it had been believed that an
82 percent charge of CFC-12 mimics a 100 percent charge of Halon
1301, recently it has become apparent that the hydraulics of the two
chemicals differ as they move through the pipes of the flooding system.
Second, and more important, CFC-12 at high concentrations has an
anesthetic effect. To properly test a flooding system a gas concentra-
tion of 5 percent must be held in the room for 10 minutes. Because
service workers have complained of headaches from CFC-12-filled
rooms, some are moving toward Halon 1301 in spite of its much higher
price.
CURRENT AND FUTURE REPORTING COUNTRY USE
In the United States, growth in new systems has remained high, an
estimated 20 percent between 1984 and 1985. For the base projection
we assume that rapid growth will continue at rates averaging 15 per-
cent through 1988. By then, many existing computers will be protected
by total flooding systems and growth is projected to decline to an aver-
age of 7 percent annually until 2000. In the base projection, the aver-
age system charge is also projected to decrease, from 320 kilograms in
1984 to 180 kilograms by 2000, as the systems are placed in progres-
sively smaller installations (partly as a result of increasing miniaturiza-
tion of computers and other electronic equipment). Combining these
effects and the implied testing, servicing, and loss projections implies
annual U.S. production of 10,200 mt by 2000.
Table 9.2 shows the U.S. baseline projections together with our esti-
mates of reporting country Halon 1301 use. Relying on industry
sources, we use a base projection for total reporting country use twice
that in the United States.
These estimates and projections are subject to considerable uncer-
tainty. Halon total flooding systems are a substantial improvement
-------
71
Table 9.2
ESTIMATED CURRENT AND PROJECTED
REPORTING COUNTRY USE OF HALON 1301
(In mt)
Year
1985
1990
1995
2000
United
States
5,400
6,700
8,300
10,200
Other
Reporting
Countries
5,400
6,700
8,300
10,200
Total
Reporting
Countries
10,800
13,300
16,500
20,400
over earlier fire-protection systems and it is difficult to estimate how
widely they will be used. The average charge is subject to significant
uncertainty depending on the types of future installations. Increased
use of the Halon for testing could lead to substantially greater chemical
use. To account for these factors we suggest a range of uncertainty
between 0.6 and 1.67 times the base levels, for both the United States
and abroad.
Estimated current use of Halon 1211, and its base projection, are
shown in Table 9.3. Halon 1211 is used primarily in hand-held extin-
guishers. According to industry sources worldwide use is about equal
to that of Halon 1301, but the United States accounts for only one-
quarter of the total. For our base projection we assume that future use
Table 9.3
ESTIMATED CURRENT AND PROJECTED
REPORTING COUNTRY USE OF HALON 1211
(In mt)
Year
1985
1990
1995
2000
United
States
2,700
3,300
4,100
5,100
Other
Reporting
Countries
8,100
10,000
12,400
15,300
Total
Reporting
Countries
10,800
13,300
16,500
20,400
-------
72
will grow at the same rate as Halon 1301 and suggest a range of uncer-
tainty between 0.6 and 1.67 times the base projected use.
United States use and reporting country use of Halons 1211 and
1301 are summarized in Fig. 9.1. The projected baseline growth is
strong but the range of uncertainty includes the possibility of almost
no growth.
-------
CO
"O
c
CO
co
13
o
CD
CO
13
"ro
3
c:
c
NOTE Historical reporting country Halon 1301 and 1211 and U S Halon 1211 use not available
Reporting Country
Halon 1301,
Halon 1211
U.S. Halon 1301
U.S. Halon 1211
1976
2000
Fig. 9.1—Estimated historical and projected U.S. and reporting country use
of Halon 1301 and Halon 1211
CO
-------
X. USE OF POTENTIAL OZONE DEPLETERS IN
THE COMMUNIST COUNTRIES
Available data on use of potential ozone-depleting substances in the
communist countries are extremely limited. The only published esti-
mates are for Soviet Union use of CFC-11 and CFC-12 from 1968
through 1975. However, we believe that the Soviet Union accounts for
the majority of use in these countries. Our approach to estimating
base projections for CFC-11 and CFC-12 is to estimate total use by
extrapolating from these historical data. Since we have little informa-
tion on the pattern of CFC uses, we do not explicitly estimate com-
munist country use by application.
The estimates of current solvent use are similarly uncertain. Car-
bon tetrachloride use is based on estimated CFC-11 and CFC-12 use.
The methyl chloroform estimate assumes that communist country use
relative to GNP is the same as the corresponding ratio in the reporting
countries, whereas the CFC-113 estimate is based on industry sources
who suggest that communist use of this solvent is quite small. Since
we have no information that Halon 1301 and 1211 are even used in the
communist countries, and because their use in the reporting countries
is so small, we assume that use in the communist countries is negligi-
ble.
Our lack of information on potential ozone-depleter use in the com-
munist countries is too great to treat the structure of uncertainty in
detail. Our uncertainty ranges are based on our ranges for use of each
chemical in the CMA reporting countries excluding the United States.
To reflect our greater uncertainty about use in the communist coun-
tries, the uncertainty ranges are constructed so that the variance of
use, conditional on the level of the GNP, is 1.25 times greater than the
corresponding variance for the non-U.S. reporting countries. As
described in Sec. II, our uncertainty range for communist country GNP
is also wider than for the reporting countries.
CFC-11 AND CFC-12 USE
Table 10.1 presents estimated CFC-11 and CFC-12 production in the
Soviet Union from 1968 through 1975 and base projections for the
communist countries from 1985 to 2000. The production estimates
were originally published in a Soviet journal (Borisenkov and Kazakov,
74
-------
75
1980) and are reprinted in the CMA reports. Reported growth was
very strong: 27 and 18 percent average annual growth rates for CFC-
11 and CFC-12, respectively. Such high rates are unlikely to have been
sustained in later years. For the period 1975 to 1985 we assume that
use grew at rates 10 percent less than the historical rates, 17 and 8 per-
cent per year, respectively. In addition, we assume that total commun-
ist country use is 15 percent greater than use in the Soviet Union (the
same assumption as CMA used in their 1983 report). These assump-
tions lead to estimated 1985 communist country use of 41,500 and
78,700 mt of CFC-11 and CFC-12. These estimates are roughly con-
sistent with estimates by other sources but are clearly more uncertain
than our estimates for other regions.
Beyond 1985 we project CFC-11 to grow at an average rate of 6 per-
cent through the end of the century and CFC-12 at an average rate of 5
percent. These rates are 2.0 and 1.67 times the base projected GNP
growth rate (3.0 percent). Such high rates are justified by the level of
development of the communist economies, a level at which uses of
CFC-containing products are likely to grow rapidly, and by the histori-
cally strong growth. We suspect that CFC-11 will grow more rapidly
Table 10.1
ESTIMATED HISTORICAL SOVIET UNION
AND PROJECTED COMMUNIST COUNTRY
USE OF CFC-11 AND CFC-12
(In mt)
Year
CFC-11
CFC-12
Soviet Union
1968
1969
1970
1971
1972
1973
1974
1975
1,400
2,200
2,500
2,900
3,700
4,100
6,200
7,500
9,800
12,200
13,500
16,200
18,500
20,100
25,900
31,700
Communist Countries
1985
1990
1995
2000
41,500
55,500
74,200
99,400
78,700
100,400
128,200
163,600
-------
76
than CFC-12 because the estimated 1985 use of CFC-12 is so much
higher than that of CFC-11 (compared with the pattern in other
regions) and because our analysis of specific applications suggests that
CFC-11 use will grow more rapidly than CFC-12 use in other world
regions. Estimated historical and projected use is depicted in Fig. 10.1.
SOLVENT USE
Table 10.2 reports our estimates of current and projected solvent use
in the communist countries. Because we have no information to the
contrary, and because CMA reporting country use of Halon 1301 and
Halon 1211 is small, we assume that communist country use of these
chemicals is negligible.
Current use of CFC-113 in the communist countries is believed to be
modest at best, although some is likely to be used in manufacturing
military and scientific electronic equipment. We estimate a nominal
5,000 mt. The baseline projection is based on projected growth in the
non-U.S reporting countries: The difference between the baseline
CFC-113 and GNP growth rates is the same as for these countries.
Thus the baseline projection for the communist countries is 6.6 percent
for the next 15 years, 3.6 percent higher than the projected GNP
growth rate. The range of uncertainty is similarly based on the range
for the non-U.S. reporting countries. Combined with uncertainty about
the level of the GNP, the range is from 0.62 to 1.62 times the base pro-
jection.
Since methyl chloroform is a popular general purpose solvent in the
reporting countries, we suspect it is also widely used in the communist
countries. We estimate current use by assuming that the ratio of
methyl chloroform use to GNP is the same as for the reporting coun-
tries. Thus communist use is estimated as 16 percent of world use, or
87,000 mt. In the reporting countries other than the United States,
methyl chloroform is projected to grow at the same rate as the GNP.
We adopt the same assumption for the communist countries, implying
a baseline projected rate of 3 percent annually. The corresponding
range of uncertainty is from 0.68 to 1.47 times the base projection in
2000.
The estimated current and projected use of carbon tetrachloride is
based on the CFC-11 and CFC-12 estimates, using the same assump-
tions about other uses as for the reporting countries. Estimated
current use is 162,700 mt, with a projected baseline growth rate of 5.2
percent. The range of uncertainty is between 0.68 and 1.47 times the
projected base level.
-------
»*—
o
V)
O
*^
§
c
1996
2000
Fig. 10.1—Estimated historical and projected use of CFC-11 and CFC-12 in the communist countries
-------
78
Table 10.2
ESTIMATED CURRENT AND PROJECTED COMMUNIST
COUNTRY USE OF CFC-113, METHYL CHLOROFORM,
AND CARBON TETRACHLORIDE
(In mt)
Year
1985
1990
1995
2000
CFC-113
5,000
6,900
9,500
13,000
Methyl
Chloroform
87,000
100,900
116,900
135,500
Carbon
Tetrachloride
159,000
205,900
267,000
346,400
-------
XI. CONCLUSIONS
We are now in a position to aggregate all of the details of the
preceding sections into global projections. This report examines the
current and likely future global production levels of the seven most
important potential ozone depleters—CFC-11, CFC-12, CFC-113, car-
bon tetrachloride, methyl chloroform, Halon 1211, and Halon 1301. It
starts by looking at the products in which these chemicals are used in
the United States and the other CMA reporting countries. It asks how
demand for these products and hence derived demand for the chemicals
themselves is likely to change over the next 15 years. It then posits
similar, but much less detailed and confident, demand trends for the
communist countries. This product-based or "bottom-up" approach
cannot capture the full demand for CFC-11 and CFC-12; hence the
report examines the magnitude of the uses not accounted for by the
method and develops a way to project these unallocated uses into the
future. By aggregating projections of demand for each chemical in the
United States, other reporting countries, and communist countries, as
well as projections of the likely shortfalls for CFC-11 and CFC-12, we
can develop global projections for each chemical. This section reports
those global projections and suggests some directions for future work.
UNITED STATES AND GLOBAL PROJECTIONS
Tables 11.1 and 11.2 summarize estimated current and projected
U.S. and total world use of the seven potential ozone depleters we have
analyzed. The tables present the range of projected use in 2000 and
the average annual growth rates from 1985 to 2000 implied by these
limits. In addition, Fig. 11.1 and Fig. 11.2 illustrate the estimated his-
torical and projected use of CFC-11 and CFC-12 in the United States
and CMA reporting countries. (Figures illustrating use of the other
chemicals are in preceding sections.) Much of the evident decline in
CFC-11 and CFC-12 use in the late 1970s was due to the discontinua-
tion of CFC use in many aerosol products in the United States and
elsewhere. As discussed in Sec. II, the bounds of the ranges of uncer-
tainty reported in the tables and figures are meant to capture a "rea-
sonable range" of outcomes for each chemical. The reader should keep
in mind that this is not an exact concept.
The projections for total CFC-11 and CFC-12 use are aggregated
from the projections for the specific applications we analyze. In
79
-------
80
Table 11.1
ESTIMATED CURRENT AND PROJECTED U.S. USE OF
POTENTIAL OZONE-DEPLETING SUBSTANCES
Use (thousands of mt)
Chemical
CFC-11
CFC-12
CFC-113
Methyl chloroform
Carbon tetrachloride
Halon 1301
Halon 1211
1985
75.0
135.0
73.2
270.0
280.0
5.4
2.7
Projected in 2000
Lower Upper
110 190
140 240
120 270
300 560
350 600
7 18
3 9
Average Annual
Growth
Lower
2.6
0.4
3.6
0.7
1.5
1.4
1.4
Rate (%)
Upper
6.3
4.0
9.1
5.0
5.2
8.3
8.3
Table 11.2
ESTIMATED CURRENT AND PROJECTED WORLD USE OF
POTENTIAL OZONE-DEPLETING SUBSTANCES
Chemical
CFC-11
CFC-12
CFC-113
Methyl chloroform
Carbon tetrachloride
Halon 1301
Halon 1211
Use
1985
341.5
443.7
163.2
544.6
1029.0
10.8
10.8
(thousands of mt)
Projected
Lower
420
460
290
630
1200
13
12
in 2000
Upper
730
830
610
1100
2100
33
32
Average
Annual
Growth Rate (%)
Lower
1.4
0.3
3.9
0.9
0.8
1.1
0.9
Upper
5.2
4.3
9.2
5.0
4.8
7.6
7.6
-------
T3
(0
W
O
0)
CO
15
C
1972
1976
1980
1984 1988
Year
1992
1996
2000
00
Fig. 11.1—Estimated historical and projected use of CFC-11 and CFC-12 in the United States
-------
800
700
600
•R 500
TJ
co
in
1
400
§
75
§ 300
200
100
1972
I ,
CFC-11
I
_L
I
J I
1976
1980
1984 1988
Year
1992
1996
00
to
2000
Fig. 11.2—Estimated historical and projected use of CFC-11 and CFC-12 in the reporting countries
-------
83
addition, we project the unallocated uses to also grow at a rate equal to
the average of the analyzed applications. As discussed in Appendix B,
the unallocated uses are likely to include substantial amounts of CFC-
12 used in refrigeration applications, and for CFC-11, unspecified
refrigeration and miscellaneous uses. In addition, the unallocated uses
include any new products that may be introduced over the next 15
years.
The U.S. and global projections are quite similar. In part this is
because the United States accounts for such a large share of global
demand and in part because our approach often infers demand outside
the United States using information about U.S. demand. As is often
true with chemicals, demand for those that are used in the largest
quantities tends to grow more slowly than demand for smaller "speci-
alty" chemicals. Of the seven chemicals, CFC-12 and methyl chloro-
form are projected to grow at the lowest rates, averaging between about
0 and 4 or 5 percent annually over the next 15 years. CFC-11 is
expected to grow slightly more rapidly, but with a comparably wide
range of uncertainty. The projected growth rates for carbon tetra-
chloride are similar to those of CFC-11 and CFC-12 because the major
use of this chemical is in producing the two CFCs. The other chemi-
cals are produced in smaller quantities and are expected to grow more
rapidly, at rates of about 4 to 9 percent for CFC-113 and 1 to 8 percent
for the Halons. Nonetheless, these growth rates are not large enough
relative to those for the chemicals used in greater quantities to change
the relative importance to potential ozone depletion of these seven
chemicals much over the next 15 years.
Keep in mind that the quantities shown in Tables 11.1 and 11.2 are
not the quantities relevant to potential ozone depletion. This is true
for two reasons. First, these are levels of use, not of emissions: Deple-
tion becomes a possibility only when these chemicals are emitted.
Most carbon tetrachloride is never emitted because it is converted into
other chemicals. Halons are rarely emitted and large amounts of
CFC-11 and CFC-12 are not emitted until years after they are initially
used. Second, emissions of equal quantities of different chemicals pose
varying levels of potential threat to stratospheric ozone. Despite its
high level of use, for example, methyl chloroform is estimated to
present only a modest threat to stratospheric ozone. CFC-11 and
CFC-12 are currently the most important potential ozone depleters and
will probably continue to be for the next 15 years.
The ranges of uncertainty reported in Table 11.1 for CFC-11 and
CFC-12 require aggregating the ranges for each application of these
chemicals. Similarly, the ranges for all of the chemicals reported in
-------
84
Table 11.2 require aggregation across world regions. These ranges were
calculated using a method described in Camm and Hammitt (1986).i
DIRECTIONS FOR FUTURE WORK
The research reported here suggests several potential directions for
future work. The first focuses on the areas where our understanding of
demand for potential ozone depleters is least complete. Demand out-
side the United States, particularly in communist countries, is an obvi-
ous candidate. Work is under way in this area, but progress is slow.
Formal data systems outside the United States are not nearly as good
as those within. Producers of these chemicals do not face the same
reporting requirements outside the United States as they do within.
And the growing policy importance of information about potential
ozone depleters 'can create incentives for governments outside the
United States, which might be asked to restrict the use or production
of these substances in the future, to make data collection more difficult
than it might otherwise be. Hence, although additional information in
this area would be beneficial, its collection is also proving to be costly.
It would also be useful to know more about the product areas or
applications of CFC-11 and CFC-12 not explicitly captured by the
"bottom-up" approach used here, in the United States and elsewhere.
As Appendix B suggests, these are likely to fall primarily in the refrig-
eration area. If they do, this is important information for policymakers
because refrigeration applications of CFCs are more difficult to displace
than other uses. Whatever accounts for the shortfalls, better informa-
tion about them would obviously improve our ability to project global
and U.S. demand for CFC-11 and CFC-12.
Finally, it would be valuable to give greater formal attention to the
sources of uncertainty underlying these projections. As explained in
Appendix C, the current formulation uses the concept of a subjective
probability distribution to combine information on different sources of
uncertainty. The details of the method are explained in Camm and
Hammitt (1986). At present, we assume that the levels of chemical use
in each application, conditional on the level of general economic
1Briefly, we assume that the growth rates for each chemical are normally distributed
and composed of one term corresponding to general economic growth and a second term
describing growth conditional on the level of GNP. The term corresponding to GNP
growth is common to all applications within a region, whereas the second term is
independent of the first and independent across applications. We assume that the
second term for the unallocated applications of CFC-11 and CFC-12 is distributed like
the average of the terms for the analyzed applications. To aggregate across regions we
assume that both' GNP and chemical use relative to GNP are correlated across regions,
both with correlation coefficients of 0.75.
-------
85
activity, are independent of one another. However, there are clearly
applications where use is positively or negatively correlated. For exam-
ple, if some of the more toxic solvents that compete with CFC-113 and
methyl chloroform were subjected to stricter regulation, both CFC-113
and methyl chloroform use could be expected to increase. Alterna-
tively, if substitution of reciprocating chillers for centrifugal units were
to occur at a more rapid pace than we project, CFC use in reciprocating
chillers would increase and use in centrifugal units would decline. We
have not been able to take these correlations into account as yet, in
part because we have too little information to estimate the size of the
effects.
In summary, we conclude that demand for potential ozone depleters
as a class is likely to grow at a modest rate over the next 15 years.
Some chemicals will grow faster than others but the relative mix of
these chemicals in the world economy should not change markedly over
this period. For specific chemicals, possibilities clearly exist for much
more rapid growth or for almost no growth at all. Further work could
allow us to refine our understanding of these basic trends.
-------
Appendix A
ESTIMATES OF CURRENT CONSUMPTION OF
CFC-11, CFC-12, METHYL CHLOROFORM, AND
CARBON TETRACHLORIDE
The estimates of current chemical consumption reported in the text
are intended to serve two purposes. First, we wish to present an accu-
rate overview of current use of each of the seven potential ozone
depleters we analyze. Second, since we base our projections of future
use on these estimates, we want estimates that fairly characterize use
in the mid 1980s. We wish to avoid basing our projections on the level
of consumption in any single year because, if that year is atypical, the
difference between consumption in that year and a more representative
estimate would be propagated in the projections for all future years and
even inflated because of the projection methodology.1
In addition, although our analysis focuses on consumption of the
potential ozone depleters, available data measure production.2 For total
use of CFC-11 and CFC-12 in the reporting countries we believe the
difference is not significant, because trade in these chemicals between
the reporting and communist countries is probably negligible. In con-
trast, U.S. use may differ significantly from U.S. production, reported
by the U.S. ITC, because of substantial imports and exports. More-
over, for both reporting country and U.S. estimates, inventory adjust-
ments across years and other factors, although probably small, could
affect the annual totals.
To accurately characterize the level of reporting country and U.S.
use of the potential ozone depleters whose production is reported, we
have developed estimates using not one year but the last several years
of reported production. These estimates are based on simple linear
regression models describing reported annual production as a function
of time, for 1979 through 1984 or 1985 if available. Our estimate of
'If two otherwise identical projections use different base levels, projected levels in all
future years will differ by the same factor as the base year levels. Because future use is
expected to be greater than current use, the absolute difference will expand over time.
2Published total production estimates are available for CFC-11 and CFC-12 in the
reporting countries and, in the United States, for CFC-11, CFC-12, methyl chloroform,
and carbon tetrachloride. For the other chemicals and regions our estimates of total con-
sumption are derived in the main text.
87
-------
88
1985 consumption is simply the level of production predicted by the
corresponding estimated regression equation. We do not use pre-1979
data because production of CFC-11, CFC-12, and carbon tetrachloride
declined sharply in earlier years, largely as a result of the ban on most
aerosol applications in the United States and other countries. The
production trends show a definite shift beginning in 1979.
This method for estimating current consumption is implicitly based
on two assumptions. First, it assumes that annual production, after
adjusting for business cycle and other fluctuations, has increased or
decreased by the same absolute amount over the period 1979 to 1985.
As will be shown below, except for a few anomalous years, this is a
good approximation. Second, the method assumes that consumption is
on average equal to production. For the reporting countries as a whole
this approximation is accurate, but for the United States the difference
between production and consumption may be wider. However, because
there is apparently no source of systematic, publicly available informa-
tion on imports and exports of these chemicals, we are unable to adjust
for either the average level of net imports or for changes in net imports
over time.
REPORTING COUNTRY CONSUMPTION OF CFC-11 AND
CFC-12
Table A.I reports the estimated regression equations for reporting
country production of CFC-11 and CFC-12. Because the 1985 CMA
data are not yet available, we base our estimates on the 1979 through
1984 reports. The regression estimates imply that reporting country
production has grown an average of about 3,000 mt annually since
1979. Table A.2 reports the actual reported production and the
amount predicted by the estimated regression equation. As shown
there, the regression predictions are quite accurate for both CFCs and
all years except 1982 and 1984. Production in 1982 was significantly
smaller than predicted by the regression equation, whereas 1984 pro-
duction was significantly larger.3 Whether 1984 marks the beginning of
a period of substantial growth or simply an aberration will not be
known for several years. For the present, we estimate 1985 production
and consumption by extrapolating along the estimated regression line
to obtain 300,000 mt of CFC-11 and 365,000 mt of CFC-12.4
3We judge the differences to be practically significant. They are not statistically sig-
nificant by conventional criteria.
4We have rounded the estimates to the nearest 5,000 mt.
-------
89
Table A.I
REGRESSION ESTIMATES OF REPORTING COUNTRY
CFC-11 AND CFC-12 PRODUCTION
CFC-11
Independent
Variable
Constant
Year
Estimated
Coefficient
-5690
3.02
t-statistic
-0.91
0.95
CFC-12
Estimated
Coefficient
-6240
3.33
t-statistic
-0.73
0.77
Dependent variable: reported annual production in thousands of mt
R-squared: 0.185 0.129
RMSE: 13.2 18.1
Sample size: 6 6
Table A.2
REPORTED AND PREDICTED REPORTING COUNTRY
CFC-11 AND CFC-12 PRODUCTION
(In thousands of mt)
CFC-11 CFC-12
Reported Predic- Differ- Reported Predic- Differ-
Year Production tion ence Production tion ence
1979
1980
1981
1982
1983
1984
1985
289.5
289.6
286.9
271.4
291.8
312.4
na
282.7
285.7
288.8
291.8
294.8
297.8
300.8
+6.8
+3.9
-1.9
-20.4
-3.0
+14.6
—
357.2
350.2
351.3
328.0
355.3
382.1
na
345.7
349.0
352.4
355.7
359.0
362.3
365.6
+11.5
+1.2
-1.1
-27.7
-3.7
+19.8
—
NOTE: na = not available.
UNITED STATES CONSUMPTION OF CFC-11, CFC-12,
METHYL CHLOROFORM, AND CARBON TETRA-
CHLORIDE
The ITC reports U.S. production of CFC-11, CFC-12, methyl chloro-
form, and carbon tetrachloride. However, because the United States
imports and exports substantial quantities of these chemicals, domestic
consumption need not correspond to production. Moreover, there have
-------
90
been major shifts in international chemical trade over the last several
years with the United States importing more and exporting less than
before.5 However, because systematic import and export data for these
chemicals are apparently not publicly available, we are unable to adjust
for this shift. As a result, our estimate of 1985 production may be a
downward-biased estimate of current consumption. However, such a
bias would not have a substantial effect on projected world consump-
tion of CFC-11, CFC-12, or carbon tetrachloride, since its only effect is
to incorrectly distribute total consumption between the United States
and other reporting countries. Since the projected growth rates for
these two regions do not differ widely, the effect on projected world
consumption would not be large. Our estimates of methyl chloroform
consumption in the non-U.S. reporting and communist countries are
based on estimated U.S. use, however. If the U.S. estimate is biased,
the estimate of the world total may also be biased.
Table A.3 reports the estimated regression equations for U.S. pro-
duction of CFC-11 and CFC-12. As shown, average annual production
has increased only about 500 mt per year over the post 1979 period. As
described above, we believe that U.S. use has grown more rapidly than
production because of increased imports and declining exports. The
low R-squared values do not indicate that the regression models do not
accurately summarize the production data. The values are low because
the general trend is essentially flat, so a linear function of time cannot
explain much of the variation in annual production.
As shown in Table A.4, the values predicted by the regression equa-
tion are quite close to the reported production levels. As for the
reporting country totals, reported 1982 production is significantly
below the model prediction and reported 1984 production is signifi-
cantly above the prediction. After rounding, our estimates for 1985
consumption of CFC-11 and CFC-12 are 75,000 and 135,000 mt, respec-
tively, about 1,000 and 7,000 mt larger than the reported 1985 produc-
tion levels.
The regression estimates for methyl chloroform and carbon tetra-
chloride are reported in Table A.5. Annual U.S. production has
declined on average over recent years, about 7,000 mt per year. As
shown in Table A.6, reported 1982 production is lower than that
predicted by the regression equation, whereas reported 1984 production
5Much of this shift is presumably attributable to the recent strength of the U.S. dol-
lar. For at least some chemicals the shift has been dramatic. For example, using data
from the Chemical Marketing Reporter profiles on trichloroethylene (January 27, 1986)
and perchloroethylene (February 3, 1986), we calculate that the increase in net imports
(imports minus exports) of these chemicals between 1980 and 1985 was about 40 and 27
percent of the 1985 U.S. consumption of these chemicals, respectively.
-------
91
Table A.3
REGRESSION ESTIMATES OF U.S. CFC-11
AND CFC-12 PRODUCTION
CFC-11
Independent
Variable
Constant
Year
Estimated
Coefficient
-1190
0.636
t-statistic
-0.49
0.52
CFC-12
Estimated
Coefficient
-474
0.307
t-statistic
-0.10
0.12
Dependent variable: reported annual production in thousands of mt
R-squared: 0.052 0.003
RMSE: 6.4 13.0
Sample size: 7 7
Table A.4
REPORTED AND PREDICTED U.S. CFC-11
AND CFC-12 PRODUCTION
(In thousands of mt)
CFC-11 CFC-12
Reported Predic- Differ- Reported Predic- Differ-
Year Production tion ence Production tion ence
1979
1980
1981
1982
1983
1984
75.8
71.7
73.9
63.5
73.0
83.9
71.8
72.4
73.0
73.7
74.3
74.9
+4.0
-0.7
+0.9
-10.2
-1.3
+9.0
133.4
133.8
147.4
117.0
134.3
152.9
134.3
134.6
134.9
135.2
135.6
135.9
-0.9
-0.8
+12.5
-18.2
-1.3
+17.0
1985 73.9 75.6 -1.7 127.9 136.2 -8.3
-------
92
Table A.5
REGRESSION ESTIMATES OF U.S. METHYL CHLOROFORM
AND CARBON TETRACHLORIDE PRODUCTION
Methyl Chloroform
Carbon Tetrachloride
Independent Estimated Estimated
Variable Coefficient t-statistic Coefficient t-statistic
Constant
Year
R-squared:
RMSE:
Sample size:
14,400
-7.13
0.400
21.1
7
1.82
-1.79
13,800
-6.80
0.233
29.2
7
1.26
-1.23
Table A.6
REPORTED AND PREDICTED U.S. METHYL CHLOROFORM
AND CARBON TETRACHLORIDE PRODUCTION
(In thousands of mt)
Methyl Chloroform
Carbon Tetrachloride
Reported
Year Production
1979
1980
1981
1982
1983
1984
1985
325.0
314.0
279.0
270.0
265.8
306.2
268.1
Predic-
tion
311.1
304.0
296.9
289.7
282.6
275.5
268.4
Differ-
ence
+13.9
+10.0
-17.9
-19.7
-16.8
+30.7
-0.3
Reported
Production
324.0
322.0
330.0
266.0
259.9
323.4
283.0
Predic-
tion
321.6
314.8
308.0
301.2
294.4
287.6
280.8
Differ-
ence
+2.4
+7.2
+22.0
-35.2
-34.5
+35.8
+2.2
is higher, the same pattern as for the CFCs in both the United States
and the reporting countries as a whole. However, reported U.S. pro-
duction of methyl chloroform and carbon tetrachloride has fluctuated
more (in absolute level) than reported CFC production, as shown by
the larger differences between the reported and predicted production
levels. After rounding, we estimate 1985 U.S. consumption of methyl
chloroform and carbon tetrachloride as 270,000 and 280,000 mt, almost
identical to the reported production levels.
-------
Appendix B
COMPARISON OF ESTIMATED TOTAL CFC-11
AND CFC-12 USE WITH OTHER SOURCES
Our estimates of total CFC-11 and CFC-12 use are constructed
"from the bottom up." That is, we estimate use in each application
considered, usually by combining estimates of total final product pro-
duced and CFC use per unit of product. Our estimate of total use is
simply the sum of the estimates for each application. In this section
we compare our estimates to total reporting country production of
CFC-11 and CFC-12 reported by the Chemical Manufacturers Associa-
tion (CMA, 1984) and total U.S. production reported by the U.S. Inter-
national Trade Commission (U.S. ITC, 1984). Our estimates are sub-
stantially lower than the CMA and ITC totals. Estimated CFC-11 use
is about 18 percent lower for the United States and 8 percent lower for
the reporting countries. For CFC-12, our U.S. estimate is 31 percent
lower and the reporting country estimate is about 23 percent too small.
To understand the source of these discrepancies we also compare our
current estimates to those in earlier Rand work (Palmer et al., 1980)
and to those published by DuPont (1978).
Our estimates for the other chemicals are based on published or
industry-supplied estimates of total use that we allocated to various
applications. These "top-down" estimates necessarily agree with the
published totals.1
COMPARISON OF ESTIMATED CFC-11 USE
Table B.I summarizes estimated current use of CFC-11 by applica-
tion and compares it to ITC and CMA reported total use in the United
States and reporting countries. The CMA estimates in each applica-
tion are for 1984, when reported production totaled 312,400 mt. The
ITC and CMA totals are based on annual production since 1979 and
are derived in Appendix A. The Rand U.S. estimate is 18 percent
lower than the adjusted ITC figure, whereas the reporting country esti-
mate is 8 percent smaller than the adjusted CMA number.
'Note that estimated carbon tetrachloride use depends largely on reported CFC pro-
duction levels and does not fall into either category.
93
-------
94
Table B.I
COMPARISON OF ESTIMATED 1985 CFC-11 USE
(In mt)
Total Reporting
United States Countries
Application
Aerosol
Foam production
Closed cell
Open cell
Refrigeration and
air conditioning
Miscellaneous
Total
"Estimated 1985
Rand
3,800
38,300
14,800
4,400
450
61,750
use is
ITC
na
na
na
na
na
75,000"
derived in
Rand
93,700
115,800
57,000
9,900
450
276,850
Appendi}
CMA
97,500
110,600
63,300
23,900
17,000
300,000"
c A. The
CMA reported applications total 312,300 mt.
The ITC does not provide estimates of use by product area. Com-
parison of the Rand and CMA estimates reveals the Rand numbers to
be comparable in aerosols and foam production but substantially lower
in the refrigeration and miscellaneous categories, 14,000 and 17,000 mt
lower, respectively. This suggests that there may be significant use in
other applications we have not explicitly considered, including some
refrigeration and air conditioning applications. The difference between
the adjusted CMA total and the Rand reporting country estimate is
23,000 mt. Comparison of the Rand U.S. and ITC estimates suggests
that 13,000 mt, about half of the unallocated uses, are in the United
States.
COMPARISON OF ESTIMATED CFC-12 USE
Estimated CFC-12 use is presented in Table B.2 together with the
ITC and CMA estimates. Again, the ITC and CMA totals are based on
recent production trends and are derived in Appendix A. The report-
ing country estimate is 23 percent smaller than the adjusted CMA fig-
ure, whereas the U.S. estimate is about 31 percent lower than the
adjusted ITC estimate. Comparison of specific product areas again
-------
95
reveals the Rand estimate to be much smaller than the CMA estimate
for refrigeration categories: We estimate only 98,000 mt in the report-
ing countries in comparison to 188,000 mt reported by CMA, a differ-
ence of 90,000 mt or 48 percent.
The Rand aerosol estimate of 115,600 mt is slightly lower than the
CMA figure of 121,300. The Rand foam production estimate is also
somewhat lower than the CMA estimate, 42,800 versus 49,200 mt,2 but
the miscellaneous use estimates are very close.
The difference between the adjusted CMA total and the Rand
reporting country estimate is 83,000 mt. The difference in the U.S.
estimates is 42,000 mt, again about half the difference in the reporting
country estimates.
Table B.2
COMPARISON OF ESTIMATED 1985 CFC-12 USE
(In mt)
United States
Application
Aerosol
Blowing agent
Closed cell
Open cell
Total
Refrigeration and
air conditioning
Mobile air conditioning
Retail food
Centrifugal chillers
Reciprocating chillers
Home appliances
Total
Miscellaneous
Total
Rand
5,700
14,900
—
14,900
50,400
4,800
1,600
300
2,400
59,500
12,700
92,800
ITC
na
na
—
na
na
na
na
na
—
na
135,000°
Total Reporting
Countries
Rand
115,600
42,800
—
42,800
73,400
9,700
3,700
1,300
10,200
98,300
25,400
282,100
CMA
121,300
30,700
18,500
49,200
na
na
na
na
na
187,500
24,100
365,000"
"Estimated 1985 use is derived in Appendix A. The CMA
reported applications total 382,100 mt.
2Note the difference in classification of foam production. Rand attributes all CFC-12
use to closed-cell foams and CMA attributes part to open-cell foams.
-------
96
COMPARISON WITH EARLIER RAND WORK
In Table B.3 we compare the Rand 1985 U.S. CFC-11 and CFC-12
nonaerosol use estimates with the 1976 estimates of Palmer et al.
(1980). Between 1976 and 1985 these applications grew at estimated
average annual rates of 5.3 percent for CFC-11 and 2.3 percent for
CFC-12.
Most of the CFC-11 growth is attributable to insulating (closed-cell)
foams. In flexible (open-cell) foam applications CFC-11 use declined
slightly reflecting an increase in the use of methylene chloride.
Estimated use of CFC-11 in chillers increased slightly and the amount
going to other applications remained approximately constant.
In the United States most CFC-12 is used in refrigeration and air
conditioning applications. Between 1976 and 1985 estimated use in
each of these decreased or remained constant, except for mobile air
conditioning where the increased number of air conditioned vehicles
more than offset the declining average charge. Estimated use in chill-
ers declined slightly, reflecting lower leakage and servicing rates in
reciprocating chillers. Estimated retail food use remained approxima-
tely constant, whereas CFC-12 use in home appliances declined
Table B.3
COMPARISON OF RAND ESTIMATES OF 1976 AND 1985
U.S. NONAEROSOL CFC-11 AND CFC-12 USE
(In mt)
CFC-11
Application
Blowing agent
Closed Icell
Open cell
Refrigeration and
air conditioning
Mobile air conditioning
Retail food
Chillers"
Home appliances
Miscellaneous
Total
1976
16,783
15,422
—
—
3,674
—
450
36,329
1985
38,300
14,800
—
—
4,400
—
450
57,950
CFC-12
1976
10,433
—
40,733
4,822
2,227
2,841
10,049
71,105
1985
14,900
—
50,400
4,800
1,900
2,400
12,700
87,100
SOURCE: 1976 estimates from Palmer et al. (1980).
"Centrifugal and reciprocating chillers combined.
-------
97
somewhat because of a reduction in the average initial charge. In con-
trast, CFC-12 use in closed-cell foam (polystyrene sheet and board) and
other uses increased substantially.
The estimates for CFC-11 and CFC-12 use in 1976 are 8,600 and
14,600 mt smaller than production reported by the U.S. ITC. In con-
trast, the current estimates are 13,250 and 42,200 mt short, indicating
a growth in unidentified uses, especially of CFC-12.3 In the next sec-
tion we attempt to identify other product areas where the CFCs might
be used.
COMPARISON WITH DUPONT ESTIMATES
Table B.4 compares 1976 CFC-11 and CFC-12 use as estimated by
Rand (Palmer et al., 1980) and by DuPont (1978). The DuPont esti-
mates for each product area include all CFCs used for that purpose,
making comparison of use by chemical difficult. For example, the total
for food processing and handling includes use of CFC-12, CFC-22, and
CFC-502. However, by making some simple assumptions we can
obtain some insight into the composition of our unallocated uses. Our
estimates of the likely uses included in our unallocated sector are sum-
marized in Table B.5.
The Rand and DuPont estimates of CFC use as a foam blowing
agent are similar. In flexible foams, Rand's estimate of CFC-11 use
exceeds DuPont's by about 3,000 mt. In rigid foams, the comparison is
complicated by the use of both CFC-11 and CFC-12. The Rand total
of CFC-11 and CFC-12, 27,200 mt, is about 1,000 mt smaller than the
DuPont total for rigid polyurethane and polystyrene foams, 28,350 mt.
However, DuPont reports an additional 2,000 mt of CFCs used in other
foam, most of which is probably CFC-11 and CFC-12. Thus DuPont
estimates greater use in rigid foam and smaller use in flexible foam
than Rand, but the total foam blowing estimates are very close.
The major differences are in the refrigeration and air conditioning
categories, especially in air conditioning and in food processing and
handling. The estimates for air conditioning, including mobile air con-
ditioning and chillers, differ by about 18,200 mt (after combining the
Rand CFC-11 and CFC-12 estimates). As indicated in note b to Table
B.4, we have all ready subtracted Rand estimates of the CFC-500 and
some of the CFC-22 included in the DuPont figure. Most of the
3Part of the difference between the estimates of unallocated uses in 1976 and 1985 is
attributable to differences in accounting procedures: Palmer et al. (1980) explicitly
accounted for exports, intermediate use, and storage, packaging, and transport losses
from the unallocated uses.
-------
98
Table B.4
COMPARISON OF RAND AND DUPONT ESTIMATES OF
U.S. 1976 NONAEROSOL CFC-11 AND CFC-12 USE
(In mt)
Rand
DuPont
Application
CFC-11 CFC-12 Quantity
CFCs
Blowing agent
Flexible foams
Rigid foams
15,422 —
12,202
11
Polyurethane
Polystyrene
Other
Total
Refrigeration and
air conditioning
Air conditioning
Food processing
and handling
Refrigerators and
freezers
Small appliances
Industrial process
refrigeration
Miscellaneous
Miscellaneous uses
Liquid fast freezing
Sterilants
Other
Total
16,783 10,433
3,674 42,960"
— 4,822
— 2,841
— —
— —
— —
— 2,722
— 5,897
450 1,430
36,329 71,105
20,094
8,255
2,041
30,390
64,819°
33,112C
3,130
1,270
2,087
907
2,722
5,897
—
184,296
11, 12
12
11, 12,
113, 114, 115
11, 12,
113, 114, 115
11, 12, 22
12, 22, 502
12
12
12, 22
12
12
12
—
All CFCs
SOURCES: DuPont (1978), Palmer et al. (1980).
'Includes CFC-12 in centrifugal and reciprocating chillers and
mobile air conditioning.
bDuPont estimate less estimated 635 mt of CFC-500 in chillers
(Palmer et al., 1980, p. 147) and 1,361, 20,865, and 13,154 mt of
CFC-22 used in in chillers, home, and supermarket air conditioning,
respectively (Palmer et al., 1980, p. 35).
°DuPont estimate less estimated 4,763 mt of CFC 502 and 680 mt
of CFC-22 used in retail food applications (Palmer et al., 1980, p.
178).
-------
99
Table B.5
ESTIMATED SHORTFALL IN RAND 1976
ESTIMATES OF U.S. CFC-11 AND CFC-12
USE COMPARED WITH DUPONT ESTIMATES
(In mt)
Shortfall
Application
Air conditioning
Food processing and handling
Refrigerators and freezers
Small appliances
Miscellaneous
Industrial process refrigeration
Total
CFC-11
9,100
—
—
—
—
—
9,100
CFC-12
—
14,150
300
1,270
907
1,044
16,671
remaining difference is likely to be CFC-11 and CFC-22. If we arbi-
trarily assume that half is CFC-11 the estimated Rand shortfall in this
category is 9,100 mt.
In food processing and handling, Rand analyzed only retail food
store refrigeration systems. As indicated in note c of Table B.4, we
have subtracted Rand estimates of CFC-22 and CFC-502 used in retail
food store applications from the DuPont figure. We do not know
whether this accounts for all the CFC-22 and CFC-502 included in this
category. Even after this adjustment, the DuPont figure exceeds the
Rand estimate by 28,300 mt. If we arbitrarily assume that CFC-12
represents half of this difference (the same share as its share of retail
food store refrigeration applications), then Rand's estimate of CFC-12
use in food handling and processing is estimated as 14,150 mt too low.
DuPont's estimate of CFC-12 use in home refrigerators and freezers
exceeds the Rand estimate by about 300 mt. Dupont reports additional
CFC-12 use of 1,270 mt of CFC-12 in small applicances and 907 mt in
miscellaneous refrigeration applications, categories that Rand did not
analyze. DuPont also reports 2,087 mt of CFC-12 and CFC-22 used in
industrial process refrigeration. If we arbitrarily assume that half this
amount is CFC-12, the implied Rand shortfall is 1,044 mt in this
category.
Table B.5 summarizes the differences as estimated above. To sum-
marize, comparison with the DuPont estimates suggests that about
9,000 mt of CFC-11 not accounted for by Rand is probably used in air
conditioning. This figure is very close to the 8,600 mt difference
-------
100
between the Rand estimate and the ITC total. In addition, some
17,000 mt of CFC-12 not accounted for by Rand may be used primarily
in food processing and handling. Again, this estimate is close to the
14,500 mt difference between the Rand and ITC totals estimates.
SUMMARY OF COMPARISONS
Rand estimates of CFC-11 and CFC-12 use in the United States and
reporting countries are smaller than ITC and CMA reported produc-
tion levels. The CFC-11 estimates are 13,000 mt lower for the United
States, and 23,000 mt lower for the reporting countries. Estimated
CFC-12 use is 42,000 mt lower for the United States and 83,000 mt
lower in the reporting countries. Thus, about half the reporting coun-
try use of each CFC that we do not account for apparently occurs in
the United States. In contrast, estimated total CFC-11 use in the
United States is about 25 percent, and CFC-12 use about 37 percent, of
the estimated reporting country totals.
Comparison of our estimates by application suggests that much of
the shortfall, particularly of CFC-12, is in refrigeration and air condi-
tioning applications. For the reporting countries, comparison with
CMA estimates suggests that our estimates fail to account for 17,000
mt of CFC-11 in miscellaneous applications and for 14,000 mt of CFC-
11 and 98,000 mt of CFC-12 used as refrigerants. Comparison with
estimates for the EEC and some other reporting countries recently fur-
nished by the European Fluorocarbon Technical Committee (EFCTC,
1985) supports this conclusion.
For the United States, Rand's current estimates are consistent with
its estimates for 1976 but the current estimates account for a smaller
share of total U.S. use (based on ITC production reports), suggesting
that use in other applications has grown. Comparison of the Rand
1976 U.S. data with DuPont estimates for the same year suggests that
the Rand estimates fail to account for about 9,000 mt of CFC-11 used
in air conditioning and 17,000 mt of CFC-12 used primarily in food
processing and handling refrigeration applications.
-------
Appendix C
DERIVATION OF SUBJECTIVE
CREDIBILITY INTERVALS
The ranges of uncertainty for projected chemical use and for each of
the three levels of uncertainty described in Sec. II can be interpreted as
subjective 80 percent credibility intervals. When one must act in the
face of uncertainty about future outcomes that will affect the desirabil-
ity of the chosen action, a long line of analytic work recommends the
use of subjective probability distributions for the relevant outcome.1
We choose to present 80 percent credibility intervals, rather than the
more conventional 90 or 95 percent levels, because experimental evi-
dence shows that people are often poor at thinking about small proba-
bilities.2 Using 80 percent credibility intervals should enable us to
assess the ranges, and readers to interpret them, more accurately.
The subjective credibility intervals for chemical use are derived from
the intervals corresponding to each level of uncertainty. The procedure
is exact if the subjective probability distributions for each level of
uncertainty are log-normal. If not, it at least provides a useful heuris-
tic for calculating the credibility intervals.
To calculate the subjective credibility interval for chemical use we
posit a random variable associated with each level of uncertainty that
is multiplied by the base projected use. These random variables are
assumed to be log-normally distributed with the median equal to one.
This assumption implies that the probability that the outcome variable
is more than Z times the projected level is equal to the probability that
it is less than 1/Z times the projected level. For example, actual GNP
is as likely to be more than 1.5 times the projected level as it is to be
less than two-thirds of that level. For projecting future chemical pro-
Raiffa (1968) for a very clear description of the principles and a historical over-
view.
For an extensive reporting of the experimental evidence on common difficulties in
evaluating probabilities see Kahneman, Slovic, and Tversky (1982). The work reported
there suggests that most people overestimate their ability to predict random variables,
especially when attempting to use high credibility-level intervals. When assessing subjec-
tive 90 or 99 percent credibility regions, the random variable falls outside the region far
more often than the theoretical 10 or 1 percent of cases.
101
-------
102
duction the log-normal distribution provides a close approximation to
most of our subjective probability distributions.3
The random variable describing the uncertainty about total chemical
use is the product of the random variables corresponding to each level
of uncertainty. Since the random variables are independent (by con-
struction) and the median of each factor is one, the base projections
reported in the text constitute the median of the subjective credibility
intervals. If, as we assume, each of the components is distributed log-
normally then so is final chemical use. Moreover, the parameters of
the distribution of chemical use can be readily obtained from the
parameters of the distribution functions of the component random
variables, which can in turn be derived from the subjective credibility
intervals. Hence, it is a simple procedure to calculate a credibility
interval for chemical use from the subjective credibility intervals for
each level of uncertainty.4
If one is unwilling to characterize his subjective uncertainty about
one or more of the random variables describing each level of uncer-
tainty as approximately log-normal, an alternative procedure can be
used. However, it does not provide an exact probability for the result-
ing credibility interval, and the bounds on that probability may be
unhelpfully wide. The procedure is to calculate the range of possible
values for chemical use obtained by setting all components at the limits
of their respective credibility intervals. For example, assume that we
believe an 80 percent credibility interval for GNP in 2000 ranges from
0.80 to 1.25 times the projected base value. Similarly, we believe that
the 80 percent credibility interval for final product use, conditional on
GNP, ranges between 0.67 and 1.50 times the projected use, and that
the 80 percent credibility interval for use of chemical Y per unit of
final product ranges from 0.50 to 2.00 times the projected use. The
alternative credibility interval for final use of chemical Y is from 0.27
(= 0.80 x 0.67 x 0.50) to 3.75 (= 1.25 x 1.50 x 2.00) times projected
use. The probability that this interval includes the actual use is at
3The ranges reported in the text are consistent with log-normal distributions in
almost all of the cases. For a few minor CFC-11 and CFC-12 applications the ranges are
not geometrically symmetric about the base projections. To calculate the ranges for total
CFC-11 and CFC-12 use we adjust the base projections for these applications to obtain
the required symmetry.
4If the distribution of X is log-normal it can be characterized by two parameters: fj.,
the expected value of the logarithm of X, and a , the variance of the logarithm of X.
These parameters can be derived from the subjective credibility interval using a table of
the normal distribution function. The parameters of the distribution of chemical use are
simply the sum of the corresponding parameters of the distributions of each of its com-
ponents, and a credibility region for any chosen significance level can be derived with
reference to the tabulated normal distribution function.
-------
103
least 0.51 (= 0.83) and no more than 0.99 (=!-(!- 0.8)3).5 In con-
trast, if these distributions are log-normal the calculated credibility
interval would include actual chemical use with probability 0.96, and
an 80 percent credibility interval would range from 0.43 to 2.30 times
the base projected use.
5In general, if the n component credibility intervals have probability levels p,, i =
1,2, . . . ,n, then the probability associated with the aggregate credibility limit will be at
least II Pj and at most 1 - II (1 - p;) (assuming the component random variables are
independent, as here).
-------
REFERENCES
Automotive News, April 24, 1985.
Borisenkov, E. P., and Y. E. Kazakov, Trudy. Glavnaia Geofizicheskaia
Obseruatoriia,, Vol. 438, pp. 62-74, 1980.
Camm, Frank, and James K. Hammitt, An Analytic Method for Con-
structing Scenarios from a Subjective Joint Probability Distribu-
tion, The Rand Corporation, N-2442-EPA, May 1986.
Chemical Manufacturers Association, Fluorocarbon Program Panel,
World Production and Sales of Chlorofluorocarbons CFC-11 and
CFC-12, various years.
Chemical Marketing Reporter, "Chemical Profile: 1,1,1-
Trichloroethane," September 27, 1982; March 28, 1983; January
20, 1986.
Chemical Marketing Reporter, "Chemical Profile: Carbon Tetra-
chloride," April 10, 1978; May 4, 1981; February 21, 1983;
February 24, 1986.
Chemical Marketing Reporter, "Chemical Profile: Fluorocarbons,"
August 7, 1978; March 10, 1986.
Chemical Marketing Reporter, "Chemical Profile: Methylene Chloride,"
February 10, 1986.
Chemical Marketing Reporter, "Chemical Profile: Perchloroethylene,"
June 18, 1979; February 3, 1986.
Chemical Marketing Reporter, "Chemical Profile: Trichloroethylene,"
January 27, 1986.
Chemical Marketing Reporter, "Fluorocarbon 11 Competitor," February
28, 1983.
Chemical Marketing Reporter, "Specialty Chemicals' Growth Pegged at
5 Pet.," May 14, 1984.
Congressional Budget Office, Economic and Budget Update, August
1985.
Council of Economic Advisors, Economic Report of the President,
February 1985.
E. I. DuPont deNemours and Company, Information Requested by EPA
on Non-Aerosol Propellant Uses of Fully Halogenated Halocarbons,
Wilmington, Delaware, March 15, 1978.
Edmonds J. A., J. Reilly, J. R. Trabalka and D. E. Reichle, An Analysis
of Possible Future Atmospheric Retention of Fossil Fuel C02,
DOE/OR/21400-1, U.S. Department of Energy, Washington D.C.,
September 1984.
105
-------
106
European Fluorocarbon Technical Committee (EFCTC), Conseil Euro-
peen des Federations de L'Industrie Chimique, Halocarbon Trend
Study 1983-1995, Brussells, 1985.
International Bank for Reconstruction and Development, World
Development Report 1984, Oxford University Press, New York,
1984.
Kahneman, Daniel, Paul Slovic, and Amos Tversky (eds.), Judgment
Under Uncertainty: Heuristics and Biases, Cambridge University
Press, Cambridge, 1982.
Manufacturing Chemist and Aerosol News, "Strong Growth Is Forecast
for Electronics Chemicals," August 1985, p. 25.
Mining Journal, "Demand Surge Forecast for Electronic Metals," July
5, 1985, p. 5.
Mobay Chemical Corporation, Urethane Market Summary,
Polyurethane Division, 1982-1985.
Modern Plastics (various issues).
Mooz, W. E., S. H. Dole, D. L. Jaquette, W. H. Krase, P. F. Morrison,
S. L. Salem, R. G. Salter and K. A. Wolf, Technical Options for
Reducing Chlorofluorocarbon Emissions, The Rand Corporation,
R-2879-EPA, March 1982.
National Academy of Sciences, Causes and Effects of Changes in Stra-
tospheric Ozone: Update 1983, Washington D.C., 1984.
National Academy of Sciences, Causes and Effects of Stratospheric
Ozone Depletion: An Update, Washington D.C., 1982.
National Academy of Sciences, Halocarbons: Effects on Stratospheric
Ozone, Washington D.C., 1976.
National Academy of Sciences, Protection against Depletion of Stratos-
pheric Ozone, Washington D.C., 1979.
Nihon Reito Kucho Kogyo Kai (Japan Refrigeration and Air Condi-
tioning Association), Reito to Kucho (Refrigeration and Air Condi-
tioning), February 1985.
Palmer, Adele R., William E. Mooz, Timothy H. Quinn, and Kathleen
A. Wolf, Economic Implications of Regulating Chlorofluorocarbon
Emissions from Nonaerosol Applications, The Rand Corporation,
R-2524-EPA, June 1980.
Quinn, Timothy, Kathleen A. Wolf, William E. Mooz, James K. Ham-
mitt, Thomas W. Chesnutt, and Syam Sarma, Projected Use,
Emissions, and Banks of Potential Ozone Depleting Substances,
The Rand Corporation, N-2282-EPA, January 1986.
Raiffa, Howard, Decision Analysis: Introductory Lectures on Choices
under Uncertainty, Addison-Wesley Publishing Co., Reading,
Massachusetts, 1968.
-------
107
Ramanthan, V., R. J. Cicerone, H. B. Singh, and T. J. Kiehl, "Trace
Gases and Their Potential Role in Climatic Change," Journal of
Geophysical Research, Vol. 90, No. D3, June 20, 1985.
Reilly, John, Rayola Dougher, and Jae Edmonds, Determinant of Global
Energy Supply to the Year 2050, Institute for Energy Analysis,
Washington, B.C., 1981.
Seidel, Stephen, and Dale Keyes, Can We Delay a Greenhouse Warm-
ing?, U.S. Environmental Protection Agency, EPA-
230-10-84-001, Washington, B.C., 1983.
Stinson, Stephen C., "Chemicals for Electronics: New Growth in Com-
petitive Field," Chemical and Engineering News, August 8, 1983.
Taya, Chikako, "Environmental Protection and the United Nations—
The Convention for the Protection of the Ozone Layer and the
CFC Regulation Issue," The Jurist, July 15, 1985.
Tversky, Amos, and Baniel Kahneman, "Judgement Under Uncer-
tainty: Heuristics and Biases," Science, September 1974.
U.S. Environmental Protection Agency, Environmental Risk Assess-
ment of Dichloromethane, Carbon Tetrachloride, Trichloroethylene,
Methyl Chloroform, Tetrachloroethylene and Freon-113, Office of
Toxic Substances, Washington, B.C., September 1981.
U.S. International Trade Commission, Synthetic Organic Chemicals,
1976-1984.
United Nations Statistical Office, Industrial Statistics Yearbook, 1981.
Wolf, Kathleen A., Regulating Chlorofluorocarbon Emissions: Effects on
Chemical Production, The Rand Corporation, N-1483-EPA,
August 1980.
-------
AW ANALYTJC METHOD FOR CONSTRUCTING SCENARIOS FROM A
SUBJECTIVE JOINT POSSIBILITY DISTRIBUTION"
May 1986
-------
ACKNOWLEDGEMENTS
This paper was written by Frank Camm and Jim Hamrni U: oi
RAND corporation under contract with the U.S. Eiivironmonta 1
Protection Agency.
-------
- Ill -
PREFACE
This Note is one of a series of papers written at The Rand
Corporation on policy issues associated with chemicals that could
potentially deplete ozone in the stratosphere ("potential ozone
depleters"). Stratospheric ozone is important because the ozone layer
helps shield the earth from harmful ultraviolet radiation. Increases in
ultraviolet radiation may threaten human health, speed deterioration of
certain materials, reduce crop yields, and have a wide range of
potentially important ecological effects. Atmospheric models developed
and tested over the last decade suggest that global human emissions of
potential ozone depleters may lead to chemical reactions that reduce
stratospheric ozone, thereby increasing ultraviolet radiation with its
concomitant effects. Substantial scientific uncertainty persists about
whether human emissions of these chemicals actually threaten the
stratospheric ozone layer and, if they do, whether lower ozone levels
actually threaten human health and other activities at the earth's
surface that concern policymakers. Policymakers must act in the face of
this uncertainty, however, and Rand's work is designed to help them act
with the best information available.
To that end, The Rand Corporation is developing a series of reports
addressed to analysts and policymakers responsible for policy decisions
on emissions of potential ozone depleters in the United States and
elsewhere. These documents report the results of research that includes
extensive literature reviews, interviews with knowledgeable officials
associated with the production and use of potential ozone depleters, and
formal chemical, cost, economic, and statistical analyses. The series
should also interest the much broader audience of analysts and
decisionmakers whose organizations would feel the effects of government
policies with respect to emissions of such chemicals.
Published papers in the series include the following:
-------
- IV -
A. R. Palmer, W. E. Mooz, T. H. Quinn, and K. A. Wolf, Economic
Implications of Regulating Chlorofluorocarbon Emissions from
Nonaerosol Applications, R-2524-EPA, June 1980.
A. R. Palmer, W. E. Mooz, T. H. Quinn, and K. A. Wolf, Economic
Implications of Regulating Nonaerosol Chlorofluorocarbon
Emissions: An Executive Briefing, R-2575-EPA, July 1980.
K. A. Wolf, Regulating Chlorofluorocarbon Emissions: Effects
on Chemical Production, N-1483-EPA, August 1980.
A. R. Palmer and T. H. Quinn, Economic Impact Assessment of a
Chlorofluorocarbon Production Cap, N-1656-EPA, February 1981.
A. R. Palmer and T. H. Quinn, Allocating Chlorofluorocarbon
Permits: Who Gains, Who Loses, and What Is the Cost?
R-2806-EPA, July 1981.
W. E. Mooz, S. H. Dole, D. L. Jaquette, W. H. Krase, P. F.
Morrison, S. L. Salem, R. G. Salter, and K. A. Wolf, Technical
Options for Reducing Chlorofluorocarbon Emissions, R-2879-EPA,
March 1982.
E. M. Sloss and T. P. Rose, Possible Health Effects of
Increased Exposure to Ultraviolet Radiation, N-2330-EPA, July
1985.
T. H. Quinn, K. A. Wolf, W. E. Mooz, J. K. Hammitt, T. W.
Chesnutt, and S. Sarma, Projected Use, Emissions, and Banks of
Potential Ozone-Depleting Substances, N-2282-EPA, January 1986.
F. Camm, T. H. Quinn, A. Bamezai, J. K. Hammitt, M. Meltzer, W.
E. Mooz, and K. A. Wolf, Social Cost of Technical Control
Options to Reduce the Use of Potential Ozone Depleters in the
United States: An Update, N-2440-EPA, May 1986.
J. K. Hammitt, K. A. Wolf, F. Camm, W. E. Mooz, T. H. Quinn,
and A. Bamezai, Product Uses and Market Trends for Potential
Ozone-Depleting Substances: 1985-2000, R-3386-EPA, May 1986.
W. E. Mooz, K. A. Wolf, and F. Camm, Potential Constraints on
Cumulative Global Production of Chlorofluorocarbons,
R-3400-EPA, May 1986.
-------
- v -
This Note was produced under Cooperative Agreement No.
CR811991-02-0 with the U.S. Environmental Protection Agency.
-------
- v 11 -
SUMMARY
Over the last 12 years, photochemical models of the upper
atmosphere have suggested that chlorofluorocarbons (CFCs) and several
related chemicals may reduce the concentration of stratospheric ozone.
Reducing the concentration of stratospheric ozone may increase the
quantities of ultraviolet radiation penetrating to the earth s surface,
which may harm human health, reduce crop yields, accelerate the
degradation of certain materials, and have other important adverse
effects. These chemicals, which we call "potential ozone depleters" or
PODs, are emitted to the atmosphere primarily through human activities.
As a result, changes in government policies could reduce emissions of
PODs, thereby reducing their effects on ozone and reducing the potential
negative effects mentioned above. This Note is one of a series of
publications being produced by The Rand Corporation to support the
development of better information on this policy issue.
Many uncertainties are important to the issue of potential ozone
depletion and its possible effects. It is not even certain that PODs
affect ozone or that changes in ozone concentrations have any of the
negative effects mentioned above. The Environmental Protection Agency
(EPA) has developed a large model that allows consideration of these
uncertainties and their importance in policy decisions. It is
developing scenarios to capsulize in simple illustrations important
information about the range of outcomes possible with current
uncertainty. For example, to represent uncertainty about future global
emissions of methane, the EPA uses both a high growth and a low growth
scenario. As part of its effort, The Rand Corporation is helping the
EPA characterize the uncertainties about market and technological
factors that may affect the future global production of seven PODs:
CFC-11, -12, -113, carbon tetrachloride, methyl chloroform, and Halons
1201 and 1301. This Note explains how we use information about the
uncertainties associated with each of these chemicals separately to
develop scenarios that illustrate these uncertainties jointly in a
useful way.
-------
- vin -
It is important to view the uncertainties associated with these
chemicals jointly. For example, suppose we developed high growth
scenarios for each chemical separately and then used them to construct a
high growth scenario for the seven chemicals together. Unless the
production levels of these chemicals always moved together, one would be
highly unlikely to observe high growth levels for all of them
simultaneously. Historically, they have not moved together. Hence,
such an approach would yield a scenario so unlikely as to be irrelevant
to policy considerations.
The high growth scenario chosen must be meaningful to policymakers
in the sense that it must represent some likely range of events with,
from their perspective, a "high" effect. The policymaker's perspective
in this problem is focused primarily on ozone depletion (although
chemicals discussed here probably also have an effect on climatic
change, another concern to EPA). Hence, a "high" growth scenario should
include production patterns for those chemicals that, if ozone depletion
is a serious problem, are likely to lead to high ozone depletion.
Similarly, a low growth scenario should be associated with a low level
of potential ozone depletion. Simply put, this Note provides a way to
define production scenarios for the seven chemicals examined here that
policymakers could reasonably associate with a range of likely levels of
potential ozone depletion.
The technique used here starts with subjective probability
distributions for each chemical. These are based on detailed Rand
analysis, reported elsewhere, of possible production of these chemicals
during the period 1985-2040. Uncertainty about the growth rate for each
chemical is characterized by the probability distribution of the sum of
two normal variates. The first captures uncertainty about general
economic growth, and the second captures uncertainty about the growth in
intensity of use of each chemical relative to general economic growth.
The technique then uses these growth rates to calculate a rough
proxy for the general rate of growth of these PODs. The proxy is
defined by a "score function" that weights and sums the growth rates for
the seven chemicals. The weights used reflect three important factors
that help determine how likely a chemical is to affect ozone
-------
- ix -
concentration: its annual rate of production, the fraction of its
production ultimately emitted into the atmosphere, and its ozone-
depleting potential per gram in the atmosphere. That is, the score
function simply weights growth rates to reflect their relative potential
effects on ozone depletion. Because of the way it is constructed,
uncertainty about the value of the score function is also captured by a
normal distribution.
We define scenarios in terms of quantiles of the distribution of
the score function. For example, a "high" growth scenario is associated
with the 75th percentile of the distribution; a "low" growth scenario
corresponds to the 25th percentile. Each of these corresponds to a set
of events likely to have a "high" or "low" effect relevant to the
policymaker. For each scenario, we then pick growth rates for the seven
chemicals that, when weighted and summed, yield the value of the score
function for that scenario. An infinite variety of individual growth
rates is consistent with any value of the score function; this is a
generic problem in scenario construction. We use a simple convention to
pick growth rates that treats all sources of uncertainty equally.
Table S.I illustrates the scenarios generated with this technique
by showing the production levels for the seven chemicals in selected
years under each of three scenarios.
The specific technique developed here allows us to convolute all
uncertainties about POD growth rates analytically. To do this, we
express all sources of uncertainty as normal distributions and use some
simple approximations. The same general approach could be used with a
broader range of distributional assumptions and without the
approximations used here. It would require a simulation technique like
Monte Carlo, which is substantially more costly and usually requires
other kinds of approximations. The technique developed here illustrates
the application of a general approach and provides a specific method to
implement it. Although the general approach would easily allow more
complex implementations, it is by no means clear that they would be
superior to the technique used here.
The specific technique used here would allow greater subtlety in
the specification of the subjective probability distribution that
provides its inputs. For example, the scenarios developed here assume
-------
- x -
Table S.I
SELECTED PRODUCTION LEVELS FOR SCENARIOS
BASED ON QUANTILES OF THE SCORE FUNCTION
(In thousands of metric tons)
Scenarios/Quantiles of the
Chemical
CFC-11
CFC-12
Carbon
tetrachloride
CFC-113
Methyl
chloroform
Halon 1301
Halon 1211
Year
1985
2000
2040
1985
2000
2040
1985
2000
2040
1985
2000
2040
1985
2000
2040
1985
2000
2040
1985
2000
2040
"Low"
0.25
342
498
1017
444
555
1124
1029
1391
2827
163
367
700
545
752
1517
11
17
26
11
17
31
"Medium"
0.50
342
556
1435
444
622
1606
1029
1554
4014
163
422
1091
545
844
2179
11
20
44
11
20
53
Score Function
"High"
0.75
342
619
2022
444
696
2287
1029
1736
5686
163
485
1695
545
946
3123
11
24
76
11
24
91
that the intensities of use relative to general economic activity for
any two chemicals are unrelated. As better empirical information
becomes available on the substitutability of different PODs in
consumption or their jointness in production, the technique presented
-------
- xi -
here could easily accommodate it. The intensities of use for
substitutes in consumption would be negatively correlated; those for
chemicals produced jointly would be positively correlated. Such
relationships can have significant effects on the joint distributions of
related chemicals and should ultimately be reflected in this kind of
analysis.
The specific technique can also be used to develop scenarios that
incorporate a much broader range of factors than those presented here.
Scenarios that reflect additional factors may be important because of
the correlations among economic variables. For example, if global
economic growth is high, both the rate of production for PODs and the
activities like agriculture that may potentially be affected by ozone
depletion will grow faster. Assuring that "high economic growth"
scenarios for PODs are matched with similar scenarios for the effects of
ozone depletion on crop yields will yield a larger — and more accurate--
measure of the benefit from limiting POD emissions than would be
calculated if such scenarios were not matched. Developing scenarios
that are conditional on general economic growth would allow such
matching. The technique presented here would allow us to develop such
conditional scenarios in the future.
Understanding how all of the uncertainties associated with
stratospheric ozone depletion relate to one another is a complex task.
The EPA is approaching this task by breaking it into manageable pieces,
where complex information about uncertainties can be condensed into
scenarios that effectively illustrate the breadth of the uncertainty.
The technique developed here shows a way to assure that, even though
scenarios relevant to different sources of uncertainty are developed
separately, they can still be related to the primary issue relevant to
policymakers—ozone depletion—and hence to the goals of the analysis as
a whole.
-------
- xni -
ACKNOWLEDGMENTS
John Hoffman and Stephen Seidel organized an informal workshop that
gave us an early opportunity to present the general approach used here
and get suggestions. Suggestions from Michael Gibbs, Michael Kavanaugh,
and Gary Yohe were especially helpful. James Hodges read an earlier
draft and provided detailed comments that have significantly improved
the Note. Jan Acton helped facilitate its production and review under
tight deadlines. Mary Vaiana helped prepare the presentation of
material in this Note to the Environmental Protection Agency's March
1986 workshop, "Protecting the Ozone Layer." Participants in that
workshop, particularly Toby Page, provided helpful feedback. Alyce
Shigg oversaw production of the many drafts underlying this Note and
Patricia Bedrosian edited the final draft. We thank them all and retain
responsibility for any errors that remain.
-------
- xv -
CONTENTS
PREFACE
SUMMARY vii
ACKNOWLEDGMENTS xiii
FIGURES AND TABLES xvii
Section
I. INTRODUCTION 1
II. BACKGROUND: THE NEED TO COMBINE DISTRIBUTIONS FOR POTENTIAL
OZONE DEPLETERS 4
Wanted: A Subjective Probability Distribution for Ozone
Depletion 5
Scenarios Based on a Proxy for Stratospheric Ozone
Depletion 9
III. THE DISTRIBUTION OF THE SCORE FUNCTION 15
Aggregating Chemical Use 16
Deriving Scenarios for Individual Chemicals from the Score
Function 18
IV. SUBJECTIVE MARGINAL PROBABILITY DISTRIBUTIONS FOR POTENTIAL
OZONE DEPLETERS 22
Methodology 22
Subjective Probability Distribution for the Pre-2000
Period 25
Subjective Probability Distribution for the Post-2000
Period 27
Marginal Distributions for Individual Chemical Production . 31
-------
- xvi -
V. PRODUCTION SCENARIOS BASED ON QUANTILES OF THE
SCORE FUNCTION 33
Choosing the Quantiles To Use for Scenario Development .... 33
Chemical Use Scenarios 35
VI. CONCLUSIONS 39
APPENDIX 43
REFERENCES 51
-------
- xvii -
FIGURES
2.1. Schematic view of potential causes and effects of
stratospheric ozone depletion 6
TABLES
S.I. Selected Production Levels for Scenarios Based on Quantiles
of the Score Function x
4.1. Parameters of the Subjective Joint Probability Distribution
in the Pre-2000 Period 26
4.2. Parameters of the Subjective Joint Probability Distribution
in the Post-2000 Period 31
4.3. Production Levels at Specified Quantiles of the Subjective
Marginal Probability Distributions for Individual
Chemicals 32
5.1. Chemical Weights 34
5.2. z-statistics and B Values for Distributions of the Score
Function and Components 36
5.3. Growth Rates for Scenarios Before and After 2000 37
5.4. Selected Production Levels for Scenarios Based on Quantiles
of the Score Function 38
-------
I. INTRODUCTION
Over the past 12 years photochemical models of the atmosphere have
suggested that global human-caused emissions of certain
chlorofluorocarbons (CFCs) and related chemicals may reduce the
concentration of stratospheric ozone. We refer to these chemicals as
potential ozone depleters or PODs. A sufficient reduction in
stratospheric ozone could allow substantially greater amounts of
ultraviolet radiation to penetrate to the earth's surface, potentially
causing a wide variety of detrimental effects. These effects include
significant threats to human health, reductions in crop yields,
degradation of certain materials, and other important adverse ecological
consequences.1 Scientists studying the possibility of ozone depletion
have not reached a consensus on the likelihood that these adverse
effects will occur in the foreseeable future. However, because the PODs
survive in the atmosphere for many decades after their release, current
emissions may affect stratospheric ozone concentrations well into the
next century. Consequently, if the probability that the PODs will cause
serious adverse effects is sufficiently high, policies that reduce
current human emissions of these chemicals may be warranted.
The U.S. Environmental Protection Agency (EPA) has undertaken a
major effort to study these potential effects, to characterize
systematically the uncertainties about them, and to examine the likely
effects of alternative policies to control the emissions of PODs.
Ultimately, this effort will help policymakers compare the social costs
of controlling POD emissions with the social benefits. The social costs
of controls are represented by the benefits that must be forgone if the
commercial use of these chemicals is restricted. The social benefits
are amelioration of the kinds of adverse effects mentioned above, where
possible expressed in monetary terms.
details, see National Academy of Sciences (1976, 1979, 1982,
1984) or Ramanthan et al. (1985).
-------
The Rand Corporation is participating in this effort by, among
other things, studying the market and technological factors that affect
current and future human emissions of PODs. We focus on the seven most
important of these: CFC-11, -12, and -113, carbon tetrachloride, methyl
chloroform, and Halon 1211 and 1301. Rand has developed information on
the likely sources of uncertainty associated with production of these
seven PODs (Hammitt et al., 1986; Quinn et al., 1986). Previous Rand
reports analyze the production of each chemical independently. Given
EPA's approach to convoluting uncertainties, however, it is important to
analyze the combined effect of all these chemicals on potential ozone
depletion. Thus, we must characterize the joint uncertainty about
production of the PODs. This Note explains why that is true and
develops a simple methodology for doing so.
Our method focuses on uncertainties about production over time.2
However, it is not production but ultimate emissions of the PODs that
are relevant to potential ozone depletion. At present, we calculate the
time path of emissions for each chemical production trajectory using
deterministic algorithms similar to those used in prior Rand work
(Palmer et al., 1980) and by the Chemical Manufacturers Association
(CMA). Hence, the method does not capture the effect of any
uncertainties about the relationship between production and emission of
the chemicals.3
Section II provides additional background on EPA's policy analytic
approach to potential ozone depletion and Rand's role in that approach.
It explains the conceptual basis for integrating Rand's information on
uncertainties about future chemical production into EPA's approach.
Section III describes the specific analytic method we use to convolute
uncertainties about the production of individual PODs to derive the
probability distribution for the joint production of these chemicals.
Our subjective probability distributions for production of individual
2We treat annual production and use as equivalent, since production
inventories over periods of several years are negligible.
3The character and importance of these uncertainties is an
important topic for future analysis.
-------
- 3 -
chemicals are reported in Sec. IV. These are based on Rand's past work
on future production and emissions of PODs. Section V presents
production scenarios based on the joint distribution derived from the
distributions for each chemical. Conclusions and suggested directions
for future work are presented in Sec. VI, and the computer code used to
implement our method is documented in the Appendix.
-------
BACKGROUND: THE NEED TO COMBINE DISTRIBUTIONS
FOR POTENTIAL OZONE DEPLETERS
Many uncertainties impinge on the question of the costs imposed
over the next century by continuing emissions of PODs. These range from
questions about the future emissions trajectories of PODs and of other
gases that influence ozone concentration, to questions about the effect
of a specified set of emission paths on actual ozone concentration, to
questions about the effects of greater ultraviolet radiation on human
health, crop yields, and degradation of materials, and of other
consequences of policy interest. Uncertainties about each of these
factors contribute to the uncertainty about the magnitude of the
potential threat and the effects of alternative global strategies for
controlling the emission of PODs. Moreover, the presence of this
pervasive uncertainty is an important factor in choosing an appropriate
policy, since a more flexible policy may be preferred even to
alternatives that would perform better in a more certain environment.
The EPA has designed a strategy for comparing the effects of
alternative policies in the face of these many sources of uncertainty.
The strategy first reduces the broad range of possible futures to a
fairly small set of cases and then examines the effects of alternative
policies in each case. The cases are constructed through the following
method:
• Isolate the principal factors that contribute to uncertainty.
• For each factor, select a small number of scenarios to
represent the range of uncertainty about it.
• Construct cases by taking all possible combinations of the
scenarios for the different factors.
The Rand Corporation is analyzing uncertainties about the emissions
of seven PODs. To incorporate its work into EPA's framework, Rand must
develop scenarios to reflect the nature of uncertainty about future
emissions trajectories for these PODs. In this section, we ask what
-------
kinds of scenarios would be appropriate. We begin by noting that an
explicit probability distribution for ozone depletion would be
preferable to the current use of discrete scenarios if costs allowed.
However, because of the extreme complexity of the computer simulation
models of the atmospheric chemistry involved, computation costs prevent
the use of Monte Carlo analysis to generate a distribution function for
ozone depletion. As an alternative, we suggest an analytic method to
develop scenarios that approximate the kind of information that would be
produced by the preferred focus on a subjective probability distribution
for ozone depletion.1
WANTED: A SUBJECTIVE PROBABILITY DISTRIBUTION
FOR OZONE DEPLETION
The extent of possible ozone depletion lies at the heart of EPA's
policy concerns.2 Essentially, the concern is that certain human
activities may lead to ozone depletion, which may in turn affect the
quality of human life. Figure 2.1 summarizes, in a very cursory
fashion, the EPA's view of the problem and the links that must be
understood. To understand it, start at the top of the figure and work
down. Market and technological factors and government policies, ranging
from the ban on most uses of CFCs as aerosol propellants to workplace
safety standards, currently affect POD emissions. The EPA is
lThere are two main alternative interpretations of probability--
the frequency interpretation and the subjective interpretation. In the
frequency interpretation, which is more widely understood, "the
probability of event A" is understood to mean the relative frequency of
occurrence of event A in some (invariably hypothetical) infinite
sequence of repetitions of the mechanism in question. In the subjective
interpretation, "the probability of event A" is a representation of
one's belief about the likelihood of event A's occurring. That is, the
first presumes the existence of a stochastic process that can be
observed to get empirical information about the probability of an event,
whereas the second is a formal way of presenting a subjective judgment.
In this Note, all references to probability use the second
interpretation; we provide a formal method for developing the
implications of a set of subjective judgments about the state of the
world. We thank James Hodges for emphasizing the importance of
distinguishing these interpretations to make our intentions and
discussion clearer.
2The chemicals analyzed here may also affect general climatic
change, an issue that also concerns the EPA.
-------
Market and
technological
forces
Government
policies
Human emissions
of PODs
Emissions of other
relevant trace gases,
some human
Other
factors
POTENTIAL
STRATOSPHERIC
OZONE DEPLETION
1
Effects on
human health
1
Effects on
materials
1
Effects on
crop yields
1
Other
ecological
effects
Fig. 2.1.—Schematic view of potential causes and effects
of stratospheric ozone depletion
-------
considering new government policies that could also affect emissions.
After release to the atmosphere, PODs and other trace gases diffuse and
react in a complex fashion that may be influenced by other factors such
as global temperatures and the amount of solar radiation. One of the
potential results of these complex interactions is a reduction in the
concentration of stratospheric ozone. The diffusion and chemical
reactions are simulated by computer models that produce time profiles of
estimated ozone concentrations at different altitudes and, in some
cases, latitudes. These time profiles can in turn be transformed into
estimated time profiles of the various effects shown at the bottom of
the figure, using a variety of models.
Each box has uncertainties associated with it. As noted above, the
EPA plans to account for each of the identified uncertainties by
developing scenarios to represent alternative possible outcomes
corresponding to each box. For example, it might choose three scenarios
to represent market and technological forces, eight to represent two
possibilities each for emissions of three "other" gases, four to
represent two parameterizations each of two models of the upper
atmosphere, and so on. If these numbers of scenarios were used, the
analysis would generate 96 alternative possible time profiles for
stratospheric ozone. Each of these cases would then be compounded with
scenarios that encapsulate information about uncertainties associated
with the effects of each time profile, to generate the full number of
cases to be analyzed for alternative policies.
So many sources of uncertainty are important that even a simplistic
method of allowing each source to be represented by two or three
scenarios leads to an unwieldy number of cases to use in policy
analysis. However, the structure of the problem lends itself to a
simplification that could potentially allow a much less simplistic
treatment of uncertainty. Note that all of the information from the top
half of the figure funnels through a single time profile--that for the
extent of stratospheric ozone depletion--that is the only input required
to study the remaining sources of uncertainty.3 Thus, in theory, it is
3This is a slight oversimplification. Certain factors that affect
emissions in the top part of the figure may also be important to the
-------
- 8 -
possible to develop a subjective probability distribution for the extent
of ozone depletion that summarizes all of the uncertainties in the boxes
that feed into the POTENTIAL STRATOSPHERIC OZONE DEPLETION box in Fig.
2.1 that are relevant to the analysis of the effects in the final row of
boxes.
A probability distribution for the extent of ozone depletion could
be developed by first developing subjective probability distributions
for the uncertain quantities in each of the boxes that feed into the
POTENTIAL STRATOSPHERIC OZONE DEPLETION box and using Monte Carlo
analysis to convolute these distributions. Actually, one would want a
set of such ozone-depletion distributions, each conditional on a
specified government policy. These distributions would not only
summarize the extent of the known information about the factors that may
influence depletion but would also allow the assigning of probabilities
to particular ozone-depletion scenarios. Thus, one could state that the
(subjectively assessed) probability that ozone depletion will fall
between two specified levels in a given year is x percent (conditional
on the corresponding government policy).
The difficulty with this approach lies in representing the chemical
interactions in the atmosphere in a cost-effective way. Current models
of the upper atmosphere are too complex and costly to use for more than
a limited number of cases.
This difficulty does not eliminate the usefulness of using the
concept of a probability distribution for ozone depletion. As we shall
see, it can provide a basis for developing emission scenarios used as
inputs to the atmospheric models. Ideally, we should be able to
interpret cases easily in terms of this subjective probability
distribution. For example, a "high" case should be developed from
scenarios for market and technological factors, other trace gases, and
other factors that together yield a "high" level of ozone depletion,
perhaps one consistent with the 75th percentile of the subjective
size of effects in the bottom half of the figure. For example, the
level of general economic activity could affect the size of crops or
quantities of materials that might be harmed by ozone depletion. These
kinds of dependencies can be integrated by making the distribution for
ozone depletion explicitly a function of general economic growth.
-------
- 9 -
probability distribution for ozone depletion under current government
policies. Of course, we cannot accurately estimate the subjective
probability distribution without actually running the atmospheric
models, which we cannot do often enough to generate the distribution.
Nonetheless, if we must choose scenarios for the inputs to the
atmospheric models, it makes sense to think about how to choose them in
light of what they are ultimately meant to do: Provide the kind of
information that a subjective probability distribution for ozone
depletion would provide if we could develop it directly.
SCENARIOS BASED ON A PROXY FOR STRATOSPHERIC
OZONE DEPLETION
The scenarios developed here are intended to approximate those that
would be derived from a subjective probability distribution for the
extent of ozone depletion, if one were to be developed. We focus on
growth rates for each potential ozone depleter so that a single random
variable—the growth rate—can describe a time profile for production.
We identify the main independent sources of uncertainty for the growth
rate of each chemical, designate them as "component" random variates,
and parameterize "component" distributions for these variates. The
available information on economic and technological factors that affect
growth in use of these chemicals is embodied in the distributions for
these component variates. If the production growth rates of two
chemicals are related, they will both be dependent on at least one
common component variate.
To combine the distributions for the individual chemicals, we
develop a "score function" that relates production growth rates of the
seven PODs to a scalar value that has some policy relevance. Ideally,
the score function would be a monotone transformation of some relevant
measure of ozone depletion, so that its subjective distribution could
easily be related to the distribution of ozone depletion. Because that
is not possible, we seek the best proxy that we can implement in a
simple way. We derive the subjective probability distribution for this
score function by convoluting the distributions for each of the
component variates underlying the POD production growth rates.
-------
- 10 -
We then define scenarios in terms of quantiles of the score
function. For example, an "upper limit" growth scenario might use the
95th percentile of the distribution; a "high" growth scenario might use
the 75th percentile. We then seek the quantile of the distributions for
the component random variates that, if chosen for all component variates
simultaneously, would yield the value of the score function for the
quantile used to define the scenario. For example, using the 82nd
percentile for each component variate might yield values of growth rates
that, when placed in the score function, yield its value at the 95th
percentile of its distribution.ft We then use the production growth
rates for each chemical that are consistent with these values of the
component variates to define the scenario in terms of the chemicals
themselves. The resulting set of production paths for the seven
chemicals is a scenario that can be used to generate emissions inputs
for the atmospheric models. Consider the key steps of this procedure in
turn.
Subjective Probability Distributions for Individual Chemicals
Our approach begins as a standard Monte Carlo analysis would. We
seek a set of independent probability distributions and then derive a
joint probability distribution for the seven chemicals based on these
independent distributions. The joint distribution developed in this
Note, discussed in more detail in Sec. IV, is based on the assumption
that general economic growth is one source of uncertainty that affects
the production of all chemicals. Otherwise, uncertainties about the
production of these chemicals are unrelated. The approach, however,
could easily accommodate information on other interrelationships by
adding additional component variables to the analysis.
*Note that the quantiles used for the component variates will
differ from the quantile of the score function and will be closer to
their respective medians, unless all of the component variates are
perfectly correlated.
-------
- 11 -
Score Function
The score function is designed to do two things. First, as a
practical matter, it provides a scalar value that summarizes the
information in the subjective joint probability distribution for the
individual chemicals. This single value provides a simple way to relate
scenarios to one another. Second, and equally important, the value is
designed to have policy significance. Increased production of any of
the PODs will increase the value of the score function and should also
be associated with greater ozone depletion. Hence, "high" growth
scenarios should be associated with high ozone depletion and "low"
growth scenarios with low ozone depletion. This relationship between
the score function and the extent of ozone depletion is obviously not
perfect. If it were, we would not need the atmospheric models. But it
is designed to yield scenarios that can be interpreted roughly in terms
of the likely corresponding ozone depletion.5
One way to think about the score function is as a simple tool that
policymakers can use to rank alternative sets of POD production levels.
A simple score function would be the sum of POD production in a given
year. However, this function would be inadequate because of the widely
different effects that a unit of each POD may have on the ozone
concentration. The score function we propose weights the production
levels of the chemicals, transforming them to a standard value so that a
unit of each is believed to have approximately the same effect on the
ozone.
The EPA has proposed a similar approach to ozone depletion in the
past.6 Although our analysis does not use exactly the approach EPA
proposed, the use of a score function to order policy thinking about
chemicals that policymakers do not consider to be equally dangerous is
consistent with that approach.
'Holding other factors, including the emissions of other gases,
constant.
6See U.S. EPA (1980).
-------
- 12 -
Specifically, of the many different score functions that might be
used, we use a simple weighted sum of the production growth rates for
individual chemicals. The subjectively chosen weights are intended to
reflect the chemicals' relative potential to deplete stratospheric
ozone. The ith chemical's weight is defined as
w. H (p. f. e.)/I (p. f. e.)
where
p. = the annual global production of the chemical,
f. = the fraction of production of the ith chemical that is likely
to be released to the atmosphere, and
e. = the estimated effect of an emitted kilogram of the chemical on
stratospheric ozone relative to that of CFC-11.
None of these factors can be specified with certainty. We estimate the
1985 production and use levels for each chemical based on the best
available data (see Hammitt et al., 1986.) The fraction of production
that will ultimately be released is based on detailed study of the
applications of each chemical (see Hammitt et al., 1986, and Palmer et
al., 1980). The estimated relative ozone-depletion potencies are based
on information from atmospheric models (see Quinn et al., 1986). These
relative potencies are sensitive to assumptions in the atmospheric
models; we use them only to suggest orders of magnitude for the weights.
We construct two sets of weights: one for the period 1985-2000 and
another for 2000-2040. The weight for each chemical is proportional to
the product of the chemical's annual production at the start of the
relevant period and a subjective factor designed to capture the other
two terms, f. and e.. The factors for CFC-11, -12, arid -113 are 1
because, despite their diverse uses, the majority of annual production
of each CFC is emitted relatively promptly, and each presents about the
same potential threat to ozone per kilogram. For carbon tetrachloride,
-------
- 13 -
the weight is 0.064, the estimated share of production that is emitted
(most carbon tetrachloride produced is consumed in the production of
CFC-11 and -12). Carbon tetrachloride presents about the same potential
threat per kilogram to ozone as the three CFCs. The factor for methyl
chloroform is 0.1, since atmospheric models suggest its effect on ozone
is an order of magnitude smaller than the three CFCs. Like the CFCs,
emissions are typically prompt. Finally, we use a factor of 10 for both
Halons. Their effect per kilogram on stratospheric ozone may be an
order of magnitude or more greater than that of the CFCs above. We
choose a factor at the low end of this scale to reflect the fact that
Halons are banked for long periods of time between production and
emission, and consequently should not begin to contribute to ozone
depletion until much later, and also because a large fraction of the
banked Halon 1301 may be recovered and never emitted.
The most important result of using this weighting scheme is that
the growth rates of chemicals produced in large volume, or likely to
have a larger depletion effect per gram, contribute more to the score
function than others. As a result, the score function should serve as a
proxy for the potential for ozone depletion.
From Score Function to Component Distributions
Our approach makes ozone depletion the focus of concern, even
though it is represented only by a proxy. Any value of the score
function could be produced by an infinite variety of production growth
rates for individual chemicals. That problem lies at the core of using
scenarios; no one seriously expects any one scenario to occur in the
sense that all growth rates specified in the scenario persist as
expected over the life of the scenario. Scenarios are designed to
illustrate the implications of an underlying probability distribution or
to represent some general kind of event that the probability
distribution suggests has a significant probability of occurring. We
need a simple convention to pick the single set of time profiles that
make up a scenario.
We seek scenarios that illustrate the different time profiles of
ozone depletion that production of the seven PODs might induce.
Otherwise, we are indifferent about the specific time profiles chosen
-------
- 14 -
for individual chemicals. It seems reasonable to seek a convention that
treats sources of uncertainty equally to avoid any manipulation of the
approach aimed at emphasizing one chemical over another. Accordingly,
we choose as a convention to use the same quantiles for each of the
independent component distributions. Other conventions could be chosen,
just as scenarios with similar policy implications can be defined in
different ways.
The subjective probability distribution for ozone depletion stands
at the heart of EPA policy analysis. If that distribution could be
approached directly, that would be the best path to follow. This
section presents an alternative approach to take when the best
alternative appears too costly. Our approach still focuses on the
central point of interest—ozone depletion. It seeks a method to
generate scenarios that are likely to capture and illustrate the range
of ways in which economic and technological factors relevant to the
seven PODs could affect stratospheric ozone depletion.
The approach suggested in this section is conceptual in nature; it
could be implemented in many different ways. Section III explains the
specific method we have developed to implement this concept.
-------
- 15 -
III. THE DISTRIBUTION OF THE SCORE FUNCTION
To implement our approach, we need explicit subjective probability
distributions for production growth rates and for the score function.
In this section we develop a formulation that allows us to use a closed
analytic solution to convolute the distributions for the individual PODs
and to derive the distribution of the score function. The subjective
probability distributions for the individual chemicals are described in
Sec. IV.
Our method approximates the distribution for the score function
that would be developed if all the component variates that describe the
growth rates of individual PODs were distributed normally.l Normal
distributions appear reasonable in all but a few instances, and these
can be accommodated without serious difficulty. Our approximation would
be exact if the relative production shares of each chemical remained
constant over time. These shares change in our analysis but not enough
to seriously threaten the integrity of the general results.
The general conceptual approach outlined in Sec. II would allow a
more general implementation based on standard Monte Carlo techniques
that would require neither an assumption of approximate normality nor
the kinds of approximations used here. Whether this approach would
justify the additional costs is unclear, since a Monte Carlo analysis
requires approximations when continuous distributions are approximated
by discrete ones.
The approach presented in this section allows quick, low-cost
development of scenarios without sacrificing much accuracy. We discuss
in turn the role of linear approximations in aggregating chemical use
and the closed analytic solution we use to choose chemical growth rates
that are consistent with any scenario chosen using the score function.
Equivalently, we assume that the future quantities of PODs are
distributed approximately log-normally. The correspondence between a
normally distributed growth rate and a log-normally distributed future
production level is good for modest growth rates. Let Y = Y (1 + r)t
where Y. is production in year i and r is the growth rate. Then log Y
= log YQ + t log (1 + r) = log YQ + t r for small r.
-------
- 16 -
AGGREGATING CHEMICAL USE
The growth rate for the production of each chemical is
characterized as the sum of the growth rates of general economic
activity (GNP) and of the specific intensity of use of each chemical,
defined as the level of use relative to the GNP.2 We focus on the
average annual growth rates of GNP and intensity for each chemical. Our
subjective probability distributions for the growth rates of these
components are normal distributions. Because the score function is a
linear combination of the component variables, our uncertainty about its
value is also described by a normal distribution. The parameters of the
distribution of the score function's value can be expressed as simple
functions of the parameters of the distributions of the component
variables.
We have subjective distributions for use of the seven PODs before
2000 for different world regions and, for CFC-11 and CFC-12, for product
applications. We must aggregate these distributions to derive the
distributions for world use of each chemical and similarly aggregate
across chemicals to derive the distribution for the score function. We
rely on linear approximations in making these aggregations.
First consider the aggregation of growth rates for different uses
of a chemical or for total use of a hemical in different world regions.
For example, we have developed subjective probability distributions for
CFC-12 use in aerosols, foam blowing, refrigeration, air conditioning,
and other applications (Hammitt et al., 1986). Let x. be the amount of
CFC-12 used in the ith application. Then the total amount of CFC-12, x,
is simply I x. and the rate of change in x can be related to the rate of
change of the {x.} at any instant in time in the following way:
(l/x)(dx/dt) = I [(xi/x)(l/xi)(dxi/dt)] (3.1)
2This is an approximation, but one that works well for small growth
rates. The exact relationship is (1 + r) = (1 + g)(l + i) = 1 + g +
u, + 8uir> where r, is the growth rate for the kth chemical, g is the
growth rate for GNP, and u, is the growth rate for the intensity of use
of the kth chemical relative to GNP. For small g and u , gu, = 0 and r,
= g + u, . This standard approximation is often used in economics.
-------
- 17 -
If we define r as the growth rate of CFC-12, r. as the growth rate of
the ith application of CFC-12, and a. as the share of CFC-12 used in the
ith application (a. = x./l x.), Eq. (3.1) can be reexpressed as
r = I (a± r..) (3.2)
Given our small growth rates, r. - g + u.,3 where g is the percentage
growth rate for GNP and u. is the growth rate for the intensity of use
of the ith application relative to GNP. Hence, Eq. (3.2) can be
rewritten as
r = g + Z (a± u±). (3.3)
The derivation of Eq. (3.3) incorporates two linear approximations.
First, it assumes that the growth rates of two factors affecting a
variable can simply be summed to calculate the growth rate of the
variable itself. Given the small growth rates we are considering, this
is adequate. Second, although Eq. (3.2) i& exact at any instant, it is
not exact over a discrete period of time unless all u. are equal and
hence all a. remain constant over the period. Since we use the {a.}
corresponding to the beginning of the period, we are essentially using a
Laspeyres index to approximate aggregated growth rates; this index is
adequate so long as the {a.} do not shift too much over the period of
interest.*
Similarly, to aggregate use across world regions we use a linear
formula to approximate the global growth rate. If r. is the rate of
3See footnote 2 above.
''Alternatives would include using the weights corresponding to the
end of the period (a Paasche index) or an intermediate set of weights.
As long as the correct weights do not shift much over the period, any of
these choices will produce similar results. This is a specific example
of the general problem of defining index numbers. For further
discussion, see Hirshleifer (1976) or other economics texts.
-------
- 18 -
growth in the jth region and b. is the share of global use there, the
global growth rate
j r.). (3.4)
The {b } are fairly stable as long as uses in different regions do not
grow at markedly different rates.
The second place where linear aggregation is important is in the
definition of the score function. We define the score function as
s = Z (wk rk)
= 8 + z (wk V (3-5)
k
where wk is the subjective weight for the kth POD, described in Sec. II,
and u, is the growth rate for the intensity of use of the kth POD.
DERIVING SCENARIOS FOR INDIVIDUAL CHEMICALS FROM
THE SCORE FUNCTION
If g and u, in Eq. (3.5) are normally distributed, then s is
normally distributed as well. This observation is the key not only to
convoluting uncertainties in the components into a distribution for the
score function but also to moving in the opposite direction. Once a
scenario is defined in terms of a quantile of the distribution for s, we
can use the normality of the score function and its components to find
the common quantile for the component distributions that is consistent
with this scenario.
Start by noting that the mean (m ) and variance (v ) of s can be
s s
defined in terms of the means (m, ), variances (v, ), and covariances
(v , ) of its components:
JtK.
-------
- 19 -
m = ra + I (w. m. )
s g k k'
k
vs=vg+I (wk2vk) + 2 I (wkwt vk£) (3.6)
The value of the score function at the q th quantile of its distribution
is
s(q ) = m + v '5 z(q ) (3.7)
o o o o
where z(q ) is the value of the z-statistic corresponding to the q th
s s
quantile of the standard normal distribution. Analogously, the q th
quantiles of the growth rates of GNP and the intensity of use of the ith
chemical can be defined as:
g(q ) = m + v ' z(q )
c g g <~
and (3.8)
u.(q ) =m. + v. ' z(q ).
i Mc i i c
We would like to find the value z(q ), and thus implicitly the quantile
O
q , so that if we fix g and all the u, at their q th quantiles, s will
O ix C
take the value at its q th quantile. To do this, for a given value of
s, substitute Eq. (3.8) into Eq. (3.5) and Eq. (3.6) into Eq. (3.7) and
set Eq. (3.5) equal to Eq. (3.7). Rearranging yields
z(qc) [vg'5 + z (Wi v..-5)]
= z(qs) [vg + Z (w..2 Vi) + 2 I (wk W|l vk£)]'5 (3.9)
i k>d
or
-------
20 -
z(q ) = z(q ) / 3 (3.10)
c s
where
3 =
v + I (w . v . " )
g 11
[Vg + l (W1 Vi} + 2 l (wk
k>£
Equation (3.10) allows one easily to transform quantiles of the
distribution of the score function into the corresponding quantiles of
the distributions underlying the POD growth rates. Each scenario is
based on a value of q , which yields z(q ), which in Eq. (3.10) yields
3 S
z(q ).s We know that 3 > 1 so that the quantiles for the component
distributions are closer to their medians than the corresponding
quantiles for the distribution of the score function.6 That is, the
quantiles of the component variates that correspond to a specified
subjective probability interval for the score function will span an
interval associated with a lower level of subjective probability for the
component variates. As shown in Table 5.2 below, the quantiles of the
component distributions corresponding to the 90 percent subjective
probability interval for the score function approximately span a 62
percent subjective probability interval for each component variate.
Note that we need not determine q to calculate the growth rates
relevant to any scenario. Once z(q ) is calculated from Eq. (3.10), it
can be substituted into Eq. (3.8) and the appropriate values of the
component growth rates can be calculated.
6To see this, square the expression that defines 3 in Eq. (3.10)
and subtract the denominator from the numerator. This yields an
expression
2 I [w^'V'5] + 2 I [wiW. ((vvv.)'5 - v )] -
The first term must be positive. To sign the second, note that the
subjective analog of the Pearson correlation coefficient, v../(v.v.)'
cannot exceed unity. Hence, the second term is also positive.
Therefore the numerator must exceed the denominator and 3 must exceed
one .
-------
- 21 -
Adopting normal distributions for the component growth rates and
using linear approximations to convolute uncertainties and thereby
construct distributions for aggregations of the component random
variat.es considerably simplifies our analysis. The scenarios generated
will reflect the influence of the approximations used. The alternative
is to use a simulation technique like Monte Carlo, which itself normally
requires that we approximate the underlying distributions. Given the
likely sources of error associated with the method presented here, we
believe it provides a good, quick, and low-cost method for developing
scenarios that reflect the jointness of the underlying subjective
probability distribution for production levels of the chemicals of
interest to policymakers.
-------
22 -
IV. SUBJECTIVE MARGINAL PROBABILITY DISTRIBUTIONS
FOR POTENTIAL OZONE DEPLETERS
To construct scenarios from our subjective probability distribution
for the production growth rates for the seven chemicals we are studying,
we must choose subjectively based parameter values with which to
characterize it. This section explains how we choose these parameter
values for the period 1985 to 2040. l It begins with a brief explanation
of the methodology used to develop the distribution based on information
reported in other Rand documents. It then summarizes our analysis of
the period 1985-2000. Because Hammitt et al. (1986) document our choice
of parameter values for this portion of our analysis, we simply outline
the approach used and summarize the results. Finally, the section
explains how parameters were developed for the period 2000-2040.
METHODOLOGY
The ultimate object of interest in developing the subjective
probability distribution is how the production of seven chemicals grows
over time. To examine this, we focus on production growth rates. How
exactly should we represent correlations among growth rates over time
for each chemical? And how exactly should we express relationships
among growth rates for different chemicals in any year? Specifying
these relationships completely would entail a density of detail not
warranted by the extent of our knowledge about the relationships. We
seek a simplified specification of the relationships that captures their
most important features.
*It is obviously difficult to think about events this far in the
future. We look ahead this far to accommodate the needs of atmospheric
modelers. Because the chemicals we are studying can remain in the
atmosphere for decades following their emission, atmospheric models rely
on long time series of emissions as inputs. We try to reflect the
degree of our uncertainty about the far future in our choice of
parameter values.
-------
- 23 -
Relationships Across Time
Because our primary concern is with cumulative POD production, we
focus on average annual growth rates over long periods of time.
Specifically, to choose normal distributions to represent uncertainty
about growth rates, we must choose means and variances for our
distributions for the secular growth rate for each chemical and
covariances among growth rates for different chemicals in each period.
Because of our focus on long-term average rates we need not address year-
to-year variations associated with the business cycle or temporary
market conditions. It is reasonable to ignore these events because
their influence on the variance of the average growth rate falls as the
length of the period grows and should be quite modest over the 15-year
and longer periods we consider.
We break the total period from 1985 to 2040 into two subperiods and
examine the means and covariance structure of secular growth within
each. Parameter values for the period 1985-2000 are based on analysis
in Hammitt et al. (1986), whose results reflect subjective judgments
about the range of reasonable growth rates for chemicals in different
applications. These judgments are based on a detailed analysis of
market trends and potential changes in markets, technologies, and
regulations that could affect the use of these chemicals. Parameter
values for the period 2000-2040 are based on concepts developed in Quinn
et al. (1986). That document looks at historical trends in the
relationship between chemical use and income and uses these to project a
range of use levels over the period in question. The analysis is
necessarily less detailed than that in Hammitt et al. (1986). We draw
on results from Hammitt et al. in the later period to assure that our
assumptions in the two periods are mutually consistent.
Although the structure of growth rates within each period may be
conceived as a set of means and a covariance matrix for each period, the
relationship between the periods is more difficult to represent
parametrically. We choose a relationship that carries over the implicit
assumption of a positive correlation between yearly growth rates across
time within the two periods.2 To construct a high growth scenario for
Heuristically, we can think of a growth rate as having two
-------
- 24 -
the full period, we use high growth scenarios in each period.
Similarly, low growth scenarios use low growth rates in both periods.
Relationships between Chemicals
We treat the average annual growth rate as the sum of two terms:
(1) general economic growth and (2) growth of intensity of use relative
to GNP, which includes both growth in the product markets where the
chemical is used (relative to GNP) and growth in the use of the chemical
in the manufacture of the products. We assume that sources of
uncertainty in the intensity of use of each chemical are uncorrelated
between chemicals, both within and across periods. With a better
empirical understanding of the chemicals markets, it should be possible
to identify relationships of this kind in the future. For now, however,
we assume them away.
The general economic growth component is common to all of the
chemicals and consequently creates a positive covariance among the
growth rates for all of the chemicals. Before 2000, we divide the world
into regions and specify general economic growth and chemical intensity
distributions for each; after 2000, we treat the world as a single unit.
When dealing with more than one region, we assume that growth rates in
different regions are positively correlated with one another. The next
subsection discusses this in more detail.
In sum, our method characterizes a wide variety of factors relevant
to uncertainty about the future use of these chemicals. It reflects
interrelationships across time and across chemicals. A more complicated
framework for relating growth rates could potentially capture subtleties
not represented here. However, empirical data of the type and quality
necessary even to quantify all of the details of this system in a
historical period are not currently available. Until better data are
available, a more detailed structure is difficult to justify.
components. The first is a secular component that can be represented by
a single random variate for each chemical. The second is an annual
component that requires separate random variates for each year. These
can be independent of one another over time (though not necessarily
across chemicals). The first component embodies the positive
correlation we wish to capture in our choice of a method to relate the
two periods. The second is an additional source of uncertainty that
helps explain why the variance of a mean growth rate can decrease for
longer periods.
-------
SUBJECTIVE PROBABILITY DISTRIBUTION FOR THE
PRE-2000 PERIOD
The subjective probability distribution for the period before 2000
is taken from Hammitt et al. (1986). That source develops explicit 80
percent subjective probability intervals for the use and production of
the seven PODs through the end of the century. It divides world use
into use in three major regions, and projects future use of CFC-11 and
-12 in each of the major products in which they are used. The mean
rates of growth are based on analysis of trends and industry forecasts
for each application or chemical. Although the reported mean rates are
not constant over the period for some of the uses, we have calculated
average annual rates over the period for use here.
We aggregate our distributions for growth rates across applications
of a single chemical, and across chemical use in different regions, to
obtain our distributions for world GNP and chemical intensity growth
using the methods described in Sec. III. When aggregating distributions
across regions, we use a positive correlation (a coefficient of 0.75)
across regions for both GNP and chemical intensity growth rates. The
parameters of the resulting distributions are reported in Table 4.1.
The table displays the parameters of the growth rate distributions
in two forms. The columns labeled "intensity" include the mean and
standard deviation of our distributions for the rate of growth of
intensity of use of each chemical, relative to general economic growth.
The last row of the table shows the mean and standard deviation of our
distribution for the general economic growth rate itself in these
columns. The columns labeled "Production" report the mean and standard
deviation of our distributions for growth rates for production of each
chemical, including the effects of both general economic growth and
intensity of chemical use. The mean of our distribution for the
production growth rate is the sum of the means of our distributions for
general economic growth and intensity of use. The standard deviation of
the distribution for the production growth rate is the square root of
the sum of variances for general economic growth and intensity of use.
-------
- 26 -
Table 4.1
PARAMETERS OF THE SUBJECTIVE JOINT PROBABILITY
DISTRIBUTION IN THE PRE-2000 PERIOD
(Use in thousands of metric tons, rates in "„ per year)
Intensity
Chemical
CFC-11
CFC-12
Carbon tetrachloride
CFC-113
Methyl chloroform
Halon 1301
Halon 1211
Gross global product
1985
Use
341.5
443.7
1029.0
163.2
544.6
10.8
10.8
Mean
0.02
-1.00
-0.49
3.27
-0.32
1.08
0.96
3.28
Standard
Deviation
0.98
1.05
1.02
1.67
1.10
2.29
2.33
1.15
Production
Standard
Mean Deviation [a]
3.30
2.28
2.79
6.55
2.96
4.36
4.24
1.51
1.56
1.54
1.03
1.59
2.56
2.60
[a] The covariance among the chemical production growth rates is the
variance of gross global product or 1.32.
As shown by Table 4.1, we expect use of CFC-113 and the Halons to
grow more rapidly during the remainder of the century than do the other
chemicals. This reflects a typical pattern of chemical use, in that
relatively recently marketed "specialty" chemicals that are produced in
limited quantities grow rapidly as they are adopted in applications to
which they are well suited. Older chemicals, that are produced in
larger quantities, may not grow as quickly because they have already
been adopted in the applications for which they are best suited. The
covariance among the production growth rates is equal to the variance of
general global economic growth because the intensities are uncorrelated
with one another and with general economic growth.3
3The total growth rate for the ith chemical, r., equals g + u.,
where g is the general growth rate and u. is the growth of intensity of
use of the ith chemical relative to general economic growth. Hence,
cov(r.,r.) = cov(g + u^g + u.) = var(g) + covCg.up + cov(g,u ) +
cov(u.,u.) = var(g), since all the covariances in the last step are
equal to zero.
-------
SUBJECTIVE PROBABILITY DISTRIBUTION FOR THE
POST-2000 PERIOD
The subjective probability distribution for the post-2000 period is
derived from information from a number of sources. We chose parameter
values that reflect the rate of general economic growth based on data
from William Nordhaus and Gary Yohe.h We chose parameter values to
reflect the rate of growth of chemicals relative to general economic
growth based on concepts developed in Quinn et al. (1986) and the
parameter values chosen for the pre-2000 period. Information from the
pre-2000 period is used to assure internal consistency in the
distribution. This subsection first considers how information from
these sources was used and then reports the resulting distribution for
the post-2000 period.
Development of the Subjective Probability Distribution
During the post-2000 period, it is much more difficult to rely on
detailed analyses of individual chemical applications and markets than
it is in the pre-2000 period. Our basic level of uncertainty about
events in this period makes it difficult to imagine the range of events
that might occur. Quinn et al. (1986), for example, present the best
available analysis of the kinds of events that might be important to the
markets for PODs after the turn of the century. But even this analysis
generally shows less variation in growth rates after 2000 than before.
In particular, it allows for no variation in the rate of general
economic growth. As a result, we build on the basic concepts developed
in Quinn et al. (1986), but seek a more realistically broad range of
uncertainty for the growth of production during this period.
Consider first the parameter values of the distribution for the
global economic growth rate. We develop parameters for this
distribution based on distributions Nordhaus and Yohe have developed on
the projected rates of population and labor productivity growth. We
start with a weighted average of their parameters for these
distributions over the periods 2000-2025 and 2025-2050. We posit a
""Personal communication with Gary Yohe, Wesleyan University,
Middletown, Connecticut, 8 January 1986.
-------
- 28 -
correlation coefficient of -0.2 between population growth and growth in
labor productivity over the period 2000-2040.5 From these, we calculate
parameters for use in our analysis. Since the GNP growth rate is
approximately the sum of the population and labor productivity growth
rates,6 our mean GNP growth rate is the sum of the means that Nordhaus
and Yohe use for population and labor productivity, and our variance is
the sum of the variances that they use for population and labor
productivity, less a small quantity to take account of the negative
correlation between them.7
Now consider parameter values of the distributions for growth rates
of intensity of use relative to general economic growth. The most
important concept in Quinn et al. (1986) is that the rate of growth in
the production of PODs will tend toward the rate of general economic
growth over the long run. Quinn et al. use gross national product (GNP)
per capita as a measure of general economic income. We believe a more
appropriate measure is GNP itself, since this captures growth in the
income of individuals and in the population as a whole.8 Hence, we
assume that on average intensity of use will not change over the long
run.
5Global labor productivity and population are likely to be
inversely correlated over this period because population increases are
likely to be in the poorer nations that suffer from shortages of human
and physical capital. Using an input-output model of the world economy,
Leontief (1979) simulates how variations in population are likely to be
inversely correlated with per capita GNP. His results are consistent
with the use of a correlation coefficient of -0.2.
6See footnote 2 in Sec. III.
7The adjustment term is (-0.2)(2)(standard deviation for population
growth rate)(standard deviation for growth rate of labor productivity).
This term is small enough so that the choice of a correlation
coefficient does not change our results much.
8 It has been suggested that a rate of growth between that of GNP
per capita and GNP itself might be most appropriate. That is because,
as indicated above, it is likely that population and GNP per capita are
negatively correlated. We recognize that by reflecting this negative
correlation in our distribution for GNP itself. Once that relationship
is accounted for, GNP is a more useful measure of income than GNP per
capita.
-------
For most chemicals, that implies a mean of zero for the rate of
growth of intensity. The Halons are treated slightly differently. For
Halon 1301, we expect recycle and recovery to become important after
2000, creating a source of Halon 1301 that can compete with new
production. As a result, even if the annual demand for new Halon 1301
grows at the rate of growth of the GNP, production will grow more
slowly. This implies a negative growth rate of intensity of use. For
Halon 1201, we expect penetration of the fire extinguishant market to
slow beyond 2000, leading to a slightly negative growth rate for
intensity of use.
Appropriate variances are harder to choose for these distributions
because the range of uncertainty is difficult to conceptualize in
markets for individual chemicals this far into the future. We start
with the ranges suggested in Quinn et al. (1986). This document
develops ranges of growth rates for CFC-11 and -12 and, to a lesser
extent, CFC-113. We then compare these ranges with the ranges for these
chemicals developed in Quinn et al. and Hammitt et al. (1986) to the pre-
2000 period, a period we understand much better.
Our comparison of the range of growth rates in the two periods is
based on a second important concept developed in Quinn et al.
Uncertainty in either period reflects the joint probabilities of many
individual events over time. In a short time period, the number of
events relevant to our analysis--changes in regulation, technology,
product line, and so on--is small, leading to a possibility that any one
of these could lead to very large changes relative to the mean rate of
growth. Over a longer period of time, these events tend to have
offsetting effects, suggesting that a reasonable range of average annual
growth rates in intensity of use should fall over time if the range of
events likely in any fixed length of time remains constant. This is
simply a reflection of the fact that the variance of a mean of a set of
independent, identically distributed random variables falls as the
number of variables rises. Because the pre-2000 period is so much
shorter than the post-2000 period, we must adjust variances in the two
periods before comparing them.
-------
- 30 -
We can use this view of the uncertainties underlying our
distributions for intensity growth rates to choose the variance of
growth rates in the post-2000 period based on those in the pre-2000
period. To do so, calculate the variance for the pre-2000 period by
squaring the chosen standard deviation for each chemical. Multiply this
by 15/40 = 0.375 to find the equivalent variance for the longer post-
2000 period.9 Adjust this variance up by a factor to reflect our belief
that uncertainty is higher in the post-2000 period. Choosing this
factor is inherently arbitrary; after a review of the analysis of likely
trends in the post-2000 period in Quinn et al. (1986) and other
available sources, we choose a factor of 1.25. Use this new variance as
a basis for the standard deviation of the post-2000 distribution for the
rate of growth of intensity of use relative to general economic growth.
The standard deviations reported in Table 4.2 are based on this
calculation.
Parameters of the Joint Distribution
Table 4.2 presents the parameter values of the subjective
probability distribution for growth rates in the post-2000 period. Note
that the means of our subjective probability distribution for growth
rates are generally lower than for the period before 2000, and the
standard deviations are also smaller, as discussed above. The means of
our subjective probability distribution for the growth rates for all of
the PODs except the Halons are equal. Growth rates for the Halons are
lower because we expect growth to slow as the likely new applications
9The variance of a mean of n independently and identically
distributed random variables with variance v is v/n. Hence, the ratio
of variances for means based on two groups of these variables with group
sizes n and n is n /n . We are not suggesting that events in
individual years are independent of one another. But events in a time
period as long as five years probably are more or less independent of
events in the next five-year period. The proportionality of the two
periods is the same whether we consider events within each period
associated with one-year or five-year subperiods. We emphasize here
that we are dealing with subjective distributions. We use statistical
concepts most commonly associated with a frequency view of statistics to
formalize our view of the sources of uncertainty underlying our
distributions for the two periods.
-------
- 31 -
Table 4.2
PARAMETERS OF THE SUBJECTIVE JOINT PROBABILITY
DISTRIBUTION IN THE POST-2000 PERIOD
(In °0 per year)
Intensity
Chemical
CFC-11
CFC-12
Carbon tetrachloride
CFC-113
Methyl chloroform
Halon 1301
Halon 1211
Gross global product
Mean
0
0
0
0
0
-0.45
-0.05
2.4
Standard
Deviation
0.67
0.72
0.70
1.15
0.75
1.57
1.60
0.96
Production
Mean
2.4
2.4
2.4
2.4
2.4
1.95
2.35
Standard
Deviation[a]
1.17
1.20
1.19
1.50
1.22
1.84
1.87
[a] The covariance among the chemical production growth rates
is the variance of gross global product or 0.92.
for these chemicals are exhausted. The growth is slowest for Halon 1301
where recovery and reuse of the chemical should significantly affect the
need for new production as its likely markets are penetrated. Standard
deviations are also similar for all of the PODs but CFC-113 and the
Halons. Regulatory uncertainty will remain high for CFC-113. We are
significantly less certain in general about the future of the Halons
than about the future of the other chemicals.
MARGINAL DISTRIBUTIONS FOR INDIVIDUAL CHEMICAL PRODUCTION
Table 4.3 presents some illustrative quantiles of the marginal
distributions for production of individual chemicals in 2000, 2020, and
2040, together with estimated current world use. These results suggest
that a wide range of outcomes are possible as we move into the future.
To understand the full implications of the subjective probability
distribution described here, we must view outcomes for chemicals
jointly. The joint results are presented in Sec. V.
-------
- 32 -
Table 4.3
PRODUCTION LEVELS AT SPECIFIED QUANTILES OF THE SUBJECTIVE MARGINAL
PROBABILITY DISTRIBUTIONS FOR INDIVIDUAL CHEMICALS
(In thousands of metric tons)
Chemical/
Year
CFC-11
1985
2000
2040
CFC-12
1985
2000
2040
Carbon
tetrachloride
1985
2000
2040
CFC-113
1985
2000
2040
Methyl
chloroform
1985
2000
2040
Halon 1301
1985
2000
2040
Halon 1211
1985
2000
2040
0.05
342
386
466
444
425
504
1029
1070
1280
163
262
256
545
574
672
11
11
7
11
11
8
0.25
342
479
907
444
532
1001
1029
1335
2519
163
348
605
545
721
1349
11
16
21
11
16
25
Quant ile
0.50
342
556
1435
444
622
1606
1029
1554
4014
163
422
1091
545
844
2179
11
20
44
11
20
53
0.75
342
644
2263
444
725
2565
1029
1807
6371
163
512
1956
545
986
3506
11
26
92
11
26
111
0.95
342
794
4326
444
901
4994
1029
2237
12290
163
671
4476
545
1229
6893
11
37
259
11
37
315
-------
- 33 -
V. PRODUCTION SCENARIOS BASED ON QUANTILES
OF THE SCORE FUNCTION
Applying the methods described in Sec. Ill to the subjective
probability distribution defined in Sec. IV yields scenarios that we can
use to project alternative futures for the seven PODs analyzed here.
This section explains how the final calculations are made and presents a
set of production scenarios based on our technique.
CHOOSING THE QUANTILES TO USE FOR SCENARIO DEVELOPMENT
A production scenario is based on a particular quantile of the
distribution of the score function defined in Sec. III. We use the 5th,
25th, 50th, 75th, and 95th percentiles as a basis for scenarios that
represent, respectively, lower limit, low, middle, high, and upper limit
cases relevant to policy decisions. Other scenarios could obviously be
developed without difficulty using the techniques described here. These
five appear to describe the relevant policy space in a way that
facilitates analysis.
Viewed in the context of EPA's policy analysis, the middle three
scenarios—representing low, middle, and high growth--should be the most
useful. The middle case represents a scenario defined such that the
effect of these seven PODs on ozone depletion is equally likely to be
greater or smaller than the effect corresponding to this scenario.
Analogously, the low and high growth scenarios are defined such that the
probabilities of greater or lesser effects as a result of these
chemicals, conditional on the effect being greater or smaller than the
median case, are equal. Thus, these three scenarios encapsulate
information on regions of the distribution for ozone depletion that
reflect the likely range of outcomes for these seven chemicals. The
limiting scenarios at the 5th and 95th percentiles are better used to
think about the outer bounds of reasonable results than to convolute
with other sources of uncertainty in EPA's planned generation of many
cases.
-------
- 34 -
We start the development of scenarios with chemical weights for the
score function that reflect relative production levels in 1985. Table
5.1 presents these weights. These weights, together with the values of
the standard deviations from Sec. IV and Eq. (3.8) from Sec. Ill, allow
us to calculate the value 3 with which to transform z-statistics from
the distribution for the score function into z-statistics for the
component economic-growth and intensity distributions. These identify
intensity and general economic growth rates that can be used to
calculate a growth rate for each chemical for each scenario during the
pre-2000 period. These growth rates, applied to actual production
levels in 1985, allow us to calculate the production paths relevant to
each chemical for each scenario up to 2000.
The calculations for the post-2000 period are analogous. We use
different weights and consequently a different value of (5, based on 2000
production levels for the 50th percentile growth scenario. These are
also reported in Table 5.1.
Table 5.1
CHEMICAL WEIGHTS
Chemical
CFC-11
CFC-12
Carbon tetrachloride
CFC-113
Methyl chloroform
Halon 1301
Halon 1211
1985
0.265
0.344
0.051
0.126
0.043
0.085
0.085
2000
0.255
0.285
0.045
0.193
0.038
0.092
0.092
Difference
-0.010
-0.059
-0.006
0.067
-0.005
0.007
0.007
NOTE: The concepts underlying these weights are explained in
Sees. II and III. They reflect the production level, share of
production that is emitted, and potential ozone-depletion risk
per gram in the atmosphere of each chemical. Of these factors,
we assume that only production levels differ between the periods
before and after 2000.
-------
- 35 -
Median growth rates differ enough across chemicals during the pre-
2000 period to lead to significant shifts in the chemical weights
between 1985 and 2000. This means that our weighting system is only
approximate over this period; it effectively represents the use of a
Laspeyres index, with its potential problems. Median growth rates after
2000 are equal for all chemicals but the Halons. Although we use a
constant set of weights over a significantly longer period of time after
2000, the actual weights do not shift as much over this period as they
do from 1985 to 2000.*
Applying these weights in Eq. (3.8) yields g values of 1.93 before
2000 and 1.83 afterward. These yield the z-statistics for individual
scenarios shown in Table 5.2. The results in this table make it clear
how important it is to view chemicals jointly rather than individually.
To construct scenarios for the 5th percentile of the score function, we
must use z-statistics for component distributions that are consistent
with the 18th to 20th percentiles of these distributions. A 25th
percentile scenario uses z-statistics consistent with the 36th
percentile of the component distributions. Similar adjustments apply
for higher percentile scenarios. Viewing these seven chemicals together
significantly narrows the range of growth rates represented in the
scenarios before and after 2000; the effect is slightly smaller after
2000.
CHEMICAL USE SCENARIOS
Taken together with the parameter values from Sec. IV, the
z-statistics in Table 5.2 yield the growth rates in Table 5.3 as the
bases for production scenarios. Table 5.4 shows the production levels
that result from these growth rates for 1985, 2000, 2020, and 2040.
Numbers like these calculated for the intervening years provide the
basis for calculating emission scenarios, which in turn can provide
inputs to the type of policy analysis EPA is currently pursuing. Moving
largest difference is about 0.013, for Halon 1301. Weights
corresponding to the median growth scenario in 2040 are CFC-11, 0.257;
CFC-12, 0.288; carbon tetrachloride, 0.046; CFC-113, 0.196; methyl
chloroform, 0.039; Halon 1301, 0.079; and Halon 1211, 0.095.
-------
- 36 -
Table 5.2
Z-STATISTICS AND & VALUES FOR DISTRIBUTIONS
OF THE SCORE FUNCTION AND COMPONENTS
Quantile
of the
Score
Function
0.05
0.25
0.50
0.75
0.95
& value
Score
Function
-1.645
-0.675
0
0.675
1.645
z-Statistics
Pre-2000
Components
-.854
-.351
0
.351
.854
1.93
Post-2000
Components
-.899
-.369
0
.369
.899
1.83
beyond these productions scenarios, however, takes us beyond the scope
of this Note.2
2For information on how to transform production scenarios into
emission scenarios, see Palmer et al. (1980).
-------
- 37 -
Table 5.3
GROWTH RATES FOR SCENARIOS BEFORE AND AFTER 2000
(In °0 per year)
Scenarios /Quant lies
Chemical
CFG- 11
CFC-12
Carbon
tetrachloride
CFC-113
Methyl
chloroform
Halon 1301
Halon 1211
Period
Pre-2000
Post-2000
Pre-2000
Post-2000
Pre-2000
Post-2000
Pre-2000
Post-2000
Pre-2000
Post-2000
Pre-2000
Post-2000
Pre-2000
Post-2000
0.
1.
0.
0.
0.
4.
0.
1.
0.
0.
0.
1.
-0.
1.
-0.
05
47
93
39
89
15
51
03
86
93
91
12
20
12
19
0
2
1
1
1
5
1
2
1
2
1
2
-1
2
1
.25
.54
.80
.50
.78
.56
.63
.17
.77
.03
.79
.94
.07
.94
.51
0
3
2
2
2
6
2
2
2
2
2
4
2
4
2
of the Score Function
.50
.29
.40
.27
.40
.55
.40
.96
.40
.79
.40
.07
.00
.07
.47
0.
4.
3.
3.
3.
7.
3.
3.
3.
3.
3.
5.
2.
5.
3.
75
03
00
04
02
54
18
74
03
55
01
34
92
34
39
0
5
3
4
3
8
4
4
3
4
3
7
4
7
4
.95
.12
.86
.15
.91
.97
.29
.88
.94
.64
.89
.15
.22
.15
.72
-------
- 38 -
Table 5.4
SELECTED PRODUCTION LEVELS FOR SCENARIOS
BASED ON QUANTILES OF THE SCORE FUNCTION
(In thousands of metric tons)
Chemical/
Year
CFC-11
1985
2000
2040
CFC-12
1985
2000
2040
Carbon
tetrachloride
1985
2000
2040
CFC-113
1985
2000
2040
Methyl
chloroform
1985
2000
2040
Halon 1301
1985
2000
2040
Halon 1211
1985
2000
2040
0.05
342
426
617
444
471
671
1029
1183
1701
163
300
367
545
636
897
11
13
12
11
13
14
0.25
342
498
1017
444
555
1124
1029
1391
2827
163
367
700
545
752
1517
11
17
26
11
17
31
Quant ile
0.50
342
556
1435
444
622
1606
1029
1554
4014
163
422
1091
545
844
2179
11
20
44
11
20
53
0.75
342
619
2022
444
696
2287
1029
1736
5686
163
485
1695
545
946
3123
11
24
76
11
24
91
0.95
342
723
3295
444
817
3788
1029
2032
9343
163
591
3174
545
1113
5217
11
31
162
11
31
196
-------
- 39 -
VI. CONCLUSIONS
The EPA faces a difficult policy problem. The agency needs to
understand the many uncertainties that affect the relationship between
decisions to reduce the global use of PODs and the effects that such
reduction might have on human health, materials degradation, crop
yields, and other activities of interest. We address only a small
subset of these uncertainties in this Note. But the relationship of
these uncertainties to other parts of this policy problem provides the
basic motivation. On the one hand, the problem is so complex that it
must be broken into pieces if we hope to produce useful results. On the
other, the characterization of uncertainties in any part of this problem
is likely to be more useful if it properly reflects the concerns of the
problem as a whole. We offer a way to build scenarios relevant to one
piece of the problem that relates them back to the problem as a whole.
This Note addresses uncertainties associated with production of
seven PODs. Ideally, we would like to develop a subjective probability
distribution for these chemicals and convolute the uncertainties
reflected in this distribution with other sources of uncertainty
relevant to stratospheric ozone depletion. In such an approach,
developing a probability distribution for the PODs would be one step in
a process to develop a distribution for ozone depletion itself. This
cannot be done because the cost of calculating ozone-depletion profiles
is too high to allow extensive use of the Monte Carlo methods needed to
convolute all of the uncertainties relevant to ozone depletion in a
complete and detailed way.
Calculation costs dictate that only a limited number of cases be
considered in the atmospheric models used to study ozone depletion.
Hence, we must assure that the cases considered embody as much
information as possible. That is, whether we examine it in detail or
not, a distribution for ozone depletion exists that is consistent with
our assumptions about the uncertainties associated with inputs to the
atmospheric models and with the models themselves. Since we cannot
investigate this distribution of ozone depletion directly and in detail,
-------
- 40 -
the cases we use to examine parts of the distribution should tell us
something about where they lie in the distribution and about the
probability density of the distribution in their vicinity. Our method
provides a way to relate scenarios for the future production of seven
PODs to the probability distribution for ozone depletion.
The specific method we use is simple; the concept it is based on
could be used to develop more complex methods that might be more
satisfactory. It remains for future analysis to determine how much
improvement additional complexity would allow. For now, the specific
method we offer can be thought of as an illustration of a more general
conceptual approach and a practical way to implement that approach until
a better method is developed.
Here is a quick overview of the approach. Characterize uncertainty
about the growth rates of general economic activity and intensity of use
of chemicals relative to it with independent normal probability
distributions. Choose values for the means and variances of these
distributions. Define a policy-relevant score function as a linear
combination of these growth rates. Calculate the mean and variance of
its subjective probability distribution. Define scenarios in terms of
quantiles of the distribution of the score function. Identify growth
rates in the component distributions that are compatible with the value
of the score function for the quantile defining each scenario; use a
simple convention to do this. Use the growth rates of the component
distributions for a scenario to calculate the growth rates for each
chemical in that scenario.
The key to this approach is the score function. It provides a
policy-relevant scale with which to compare alternative scenarios for
seven chemicals along a single dimension. The dimension chosen is one
that should be related to the ozone depletion likely to result from the
scenarios it is used to describe. The relationship is probably crude,
but the scale reflects the kind of simple weighting scheme that EPA
policymakers have found useful in the past to specify the relative
danger associated with different chemicals and hence the joint danger
associated with any set of production levels of these chemicals. More
complicated score functions could be considered if we were interested in
using a Monte Carlo technique to convolute uncertainties. Ironically,
-------
- 41 -
as the score function comes closer to approximating the actual joint
effect of a set of chemicals on ozone depletion, using it to develop
scenarios may become less attractive. That is because a direct approach
to the subjective probability distribution for ozone depletion becomes
more attractive, eliminating the need for developing scenarios.
Restricting our subjective distributions to normal distributions
and relying on linear approximations simplifies the analysis
considerably and in fact makes a closed analytical solution to the
convolution of uncertainties possible. Abandoning normality, or some
other parametric distributions, would give us greater freedom to reflect
uncertainties as we see them but would require the use of a simulation
technique like Monte Carlo to convolute uncertainties. Once this step
is taken, we probably no longer need to rely on linear approximations.
Simulations themselves, of course, typically require simplifying
assumptions and approximations to implement them at a reasonable cost.
Whether the approximations associated with normal distributions and
linear aggregations here induce more serious errors than the
simplifications and approximations required by a technique like Monte
Carlo is an empirical question. It deserves closer attention if this
approach is to be used often in the future.
Whether an analytical approach like that used here or a Monte Carlo
approach is taken, the interrelationships among chemicals deserve more
attention. For example, CFC-11 and -12 are produced together. Although
the proportions in which they are produced are variable, it would be
surprising if cost considerations did not induce a positive correlation
in their production rates. Alternatively, CFC-113 and methyl chloroform
are substitutes, but both are subject to a similar set of government
regulations. Changes in markets and regulations that underlie the
scenarios used here could induce either a negative or positive
correlation in their growth rates. These considerations and others like
them suggest that future efforts to build scenarios for these chemicals
should give closer attention to the relationships in intensity of use
for different chemicals. Our technique makes that simple to do.
The techniques proposed here can produce a wide variety of
scenarios. They are based on simple quantiles of the score function.
The low, middle, and high growth scenarios associated with the 25th,
-------
- 42 -
50th, and 75th percentiles of its distribution characterize the relevant
policy space and should properly represent a reasonable range of effects
of market and technological developments in the larger context of EPA's
policy analysis. Developing scenarios for other factors conditional on
the rate of general economic growth might also be worth exploring. That
is because the level of the effects of ozone depletion on human
activities like materials degradation and crop yields is likely to
depend on the general rate of economic growth: Higher chemical growth
rates will presumably have larger effects" on ozone depletion which, in
turn, will have more effect on crop yields if high economic growth has
created a demand for more crops. Such relationships may prove to be
quite important in sorting out the joint effects of different sources of
uncertainty. Our technique could accommodate scenarios conditioned on
general economic growth by calculating the moments for score functions
conditional on economic growth rates; how those growth rates would be
chosen remains a problem.
In the end, the approach taken in this Note offers a simple
solution to a complex problem. More complex solutions may well justify
their additional costs but that is not immediately clear. Additional
attention to substantive issues associated with the construction of
scenarios for application to the issue of potential ozone depletion is
likely to be more productive in the short run.
-------
- 43 -
Appendix
BASIC CODE USED TO IMPLEMENT AGGREGATIONS AND CONVOLUTION
This appendix documents four BASIC programs that we employ to
convolute the distributions corresponding to different applications of
CFC-11 and CFC-12, and to total use of each chemical. In general,
variable names follow a few rules. Those beginning with "mid" indicate
mean growth rates; those beginning with "sig" indicate standard
deviations. The suffix ".g" refers to GNP, ".int" to intensity of use,
and ".com" to total use (combined GNP and intensity effects) .
The first program is used to derive the means and standard
deviations of the distributions of the intensity growth rates for CFC-11
and CFC-12 from the information presented in Hammitt et al. (1986).
That source projects use in each major application for the United States
and the other CMA reporting countries separately. It provides a
baseline projected use for each application and region together with
factors that, when multiplied by the base projected use in 2000,
characterize an 80 percent subjective probability interval for use. The
program uses the estimated 1985 use, the projected base use in 2000, the
factors characterizing the uncertainty range, and the parameters of the
subjective distribution for GNP growth to calculate the parameters of
,the intensity growth rates for each region and application. It also
convolutes these distributions to calculate the parameters of the
distribution of rate of growth of intensity of total CFC-11 or CFC-12
use by region.
The second program is used to aggregate chemical use across
regions. It takes as input the estimated 1985 use, projected baseline
2000 use, and factors characterizing the 80 percent subjective
probability interval to calculate the parameters of the intensity growth
rates by region, and convolute these to find the parameters of the world
intensity growth rate. A similar program that uses the parameters of
the intensity growth rate distributions for each region is used to
calculate the parameters of the distribution of world intensity growth
rates for CFC-11 and -12.
-------
- 44 -
The third program calculates the parameters of the score function
distribution. As input, it requires the parameters of the distribution
of world GNP growth/ world intensity growth for each chemical, and
weights that reflect the relative ozone-depletion potency of each
chemical. We run this program twice to calculate the parameters
corresponding to each period separately.
The last program is used to calculate the production of each
chemical corresponding to different quantiles of the score function, and
to quantiles of the marginal distributions for individual chemical
growth. As input it requires the intensity and GNP growth rates for the
pre- and post-2000 periods and the values of P for each period.
Variable names ending in n.e" indicate first (early) period values
whereas names ending in ".1" indicate values in the later period. These
suffixes are sometimes combined with the suffixes mentioned earlier; for
example, ".eg" refers to GNP growth in the first period.
A listing of the program code follows:
PROGRAM 1: BASIC CODE TO CALCULATE PARAMETERS OF CFC-11 AND CFC-12
INTENSITY GROWTH RATES, BY REGION AND APPLICATION, 1985-2000
intensity
100
110
120
130
140
150
160
170
180
190
200
210
220
230
240
241
250
260
270
280
290
300
310
320
rem file cfc.bas to calculate CFC-11 and 12 in
rem for US and non-US reporting companies
open "incfc" for input as #1
input#l,n,mid.g, sig.g
for i = 1 to n
input#l,bO (i) ,bn(i) , lofact (i) ,hifact (i)
next
base = 0
for i = 1 to n
lorate(i) = (bn(i) * lofact(i) / bO (i) ) A (1/15)
lorate(i) = (lorate(i) - 1) * 100
hirate(i) = (bn(i) * hifact (i) / bO (i) ) A (1/15)
hirate(i) = (hirate(i) - 1) * 100
mid(i) = (lorate(i) + hirate(i)) / 2 - mid.g
sig(i) = (hirate(i) - lorate(i)) / (2 * 1.2816)
sig(i) = (sig(i)A2 - sig.gA2)A0.5
base = base + bO(i)
next
mid.int = 0
sig.int = 0
for i = 1 to n
share (i) = bO (i) / base
midterm(i) = share(i) * mid(i)
sigt(i) = share (i) * sig(i)
-------
- 45 -
330 sigtsq(i) = sigt(i)A2
340 mid.int = mid.int + midtenn(i)
350 sig.int = sig.int + sigtsq(i)
360 next
370 mid.com = mid.int + mid.g
380 sig.com = (sig.int + sig.gA2)A0.5
390 sig.int = sig.intA0.5
400 open "outcfc" for output as #2
410 print "mid.int = ",mid.int
420 print "sig.int = ",sig.int
430 print#2,"mid.int = ",mid.int
440 print#2,"sig.int = ",sig.int
450 print#2,"mid.g = ",mid.g
460 print#2,"sig.g = ",sig.g
470 print#2,"mid.com = "/mid.com
480 print#2,"sig.com = ",sig.com
490 print#2," share(i)";" mid(i) ";" sig(i)"
500 for i = 1 to n
510 print#2,using"####*.####";share(i),mid(i),sig(i)
520 next
530 print#2,"bO","bn","lofact","hifact","lorate","hirate"
540 for i = 1 to n
550 print#2,bO(i) ,bn(i),lofact(i),hifact(i),lorate(i),hirate (i)
560 next
PROGRAM 2: BASIC CODE TO CALCULATE PARAMETERS OF WORLD
INTENSITY GROWTH RATES FOR OTHER CHEMICALS, 1985-2000
100 rem file chm.bas to calculate world intensity for other chemicals
110 open "inchm" for input as #1
120 input#l,n
130 rho(l) = 1.00
140 rho(2) = 0.75
150 mid.g(l) = 3.0
160 sig.g(l) = 1.17
170 mid.g(2) = 3.5
180 sig.g(2) = 1.17
190 mid.g(3) = 3.0
200 sig.g(3) = 1.56
210 for i = 1 to n
220 input#l,bO(i),bn(i),lofact(i),hifact(i)
230 next
240 base = 0
250 for i = 1 to n
260 lorate(i) = (bn(i) * lofact(i) / bO(i))A(1/15)
270 lorate(i) = ((lorate(i) - 1) * 100) - mid.g(i)
280 hirate(i) = (bn(i) * hifact(i) / bO(i))A(1/15)
290 hirate(i) = <(hirate(i) - 1) * 100) - mid.g(i)
300 mid(i) = (lorate (i) + hirate(i)) / 2
310 sig.c(i) = (hirate(i) - lorate(i)) / (2 * 1.2816)
320 sig(i) = ((sig.c(i)A2) - (sig.g(i)A2))A0.5
330 base = base + bO(i)
340 next
-------
- 46 -
350 mid.int - 0
360 sig.int - 0
370 for i = 1 to n
380 shared) - bO(i) / base
390 midterm(i) = share(i) * mid(i)
400 sigt(i) = share(i) * sig(i)
410 sigtsq(i) = sigt(i)A2
420 mid.int = mid.int + midterm(i)
430 sig.int - sig.int + sigtsq(i)
440 next
450 if n - 3 then goto 480
460 crossprd - 2 * sigt(l) * sigt(2)
470 goto 490
480 crossprd - 2 * (sigt(1)*sigt(2) + sigt(1)*sigt(3) + sigt(2)*sigt(3))
490 for i - 1 to 2
500 sig.int(i) - (sig.int + rho(i) * crossprd)*0.5
510 next
520 open "outchm" for output as #2
530 print "mid.int - ",mid.int
540 print "rho,std dev"
550 for i - 1 to 2
560 print rho(i),sig.int(i)
570 next
580 printf2,nmid.int - ",mid.int
590 printf2,"rho,std dev"
600 for i - 1 to 2
610 print#2,rho(i),sig.int(i)
620 next
630 print#2,"mid(i)","sig(i)","share(i)"
640 for i = 1 to n
650 print#2,mid(i),sig(i),share(i)
660 next
670 print#2,"bO","bn","lofact","hifact","lorate","hirate"
680 for i - 1 to n
690 print#2,bO(i),bn(i),lofact(i),hifact(i),lorate(i),hirate(i)
700 next
PROGRAM 3: BASIC CODE TO CALCULATE PARAMETERS OF THE DISTRIBUTION
OF THE SCORE FUNCTION
100 rem file joint.bas to calculate joint scenarios for all chemicals
110 rem uses world gnp and world intensity distributions for each chemical
120 open "injoint" for input as #1
130 input#l,n,mid.g,sig.g
140 base = 0
150 for i = 1 to n
160 input#l,wgt(i),bO(i),mid(i),sig(i)
170 share (i) = wgt(i) * bO(i)
180 base = base + share(i)
190 next
200 mid.int = 0
210 sig.int = 0
220 beta = 0
-------
- 47 -
230 for i = 1 to n
240 shared) = share (i) / base
250 midterra(i) = shared) * mid(i)
260 mid.int = mid.int + midterm(i)
270 sigt(i) = shared) * sig(i)
280 sig.int = sig.int + sigt(i)"2
290 beta = beta + sigt(i)
300 next
310 mid.com = mid.int + mid.g
320 sig.com = (sig.int + sig.gA2)A0.5
330 sig.int = sig.infO.5
340 beta = (beta + sig.g) / sig.com
350 open "outjoint" for output as #2
360 print "mid.com = ",mid.com
370 print "sig.com = ",sig.com
380 print "mid.int = ",mid.int
390 print "sig.int = ",sig.int
400 print"mid.g= ",mid.g
410 print"sig.g= ",sig.g
420 print "beta = ",beta
430 print#2,"mid.com = ",mid.com
440 print#2,"sig.com = ",sig.com
450 print#2,"mid.int = ",mid.int
460 print#2,"sig.int = ",sig.int
470 printf2,"mid.g - ",mid.g
480 print#2,"sig.g = ",sig.g
490 print#2,"beta = ",beta
500 print#2,"mid(i)","sigd)", "share (i)"
510 for i = 1 to n
520 print#2,midd) ,sig(i),share (i)
530 next
540 print#2,"wgt(i)","bO(i)","midterm(i)","sigt(i)
550 for i = 1 to n
560 print#2,wgt(i),bO(i),midterm(i),sigt(i)
570 next
PROGRAM 4: BASIC CODE TO CALCULATE CHEMICAL PRODUCTION CORRESPONDING
TO QUANTILES OF THE SCORE FUNCTION AND TO QUANTILES OF MARGINAL
CHEMICAL USE DISTRIBUTIONS
100 rem file scenar.bas to calculate quantiles of joint and individual
110 rem chemical production distributions over time
120 rem subscript 1—CFC-11, 2—CFC-12, 3—CT, 4—CFC-113, 5—MC,
130 rem 6—Halon 1301, 7—Halon 1211, 0—Joint weighted production
140 open "inscen" for input as #1
150 dim g(5,ll),q(5,7,ll),c<5,7,ll),year(ll)
160 input#l,beta.e,beta.1
170 inputf1,mid.eg,sig.eg,mid.lg,sig.lg
180 for i = 0 to 7
190 input#l,bO(i),mid.e(i),sig.e(i),mid.l(i),sig.l(i)
200 next
210 data -1.645,-0.675,0.0,0.675,1.645
220 for j = 1 to 5
-------
- 48 -
230 read zj(j)
240 zc.e(j) = zj(j) / beta.e
250 zc.l(j) - zj(j) / beta.l
260 rate.eg(j) = mid.eg + zc.e(j) * sig.eg
270 rate.e(j,0) =mid.e(0) + zj(j) * sig.e(O)
280 for t = 0 to 3
290 g(j,t) = (1 + rate.eg(j)7100)A(t * 5)
300 q(j,0,t) = bO(0) * (1 + rate.e(j,0)/100)A(t * 5)
310 next t
320 rate.lg(j) = mid.lg + zc.l(j) * sig.lg
330 rate.l(j,0) -mid.1(0) + zj(j) * sig.l(O)
340 for t = 4 to 11
350 tm3 = t - 3
360 g(j,t) - g(j,3) * (1 + rate.lg(j)7100)A(tm3 * 5)
370 q(j,0,t) = q(j,0,3) * (1 + rate.1(j,0)7100)A(tm3 * 5)
380 next t
390 for i = 1 to 7
400 rate.e(j,i) = mid.e(i) + zc.e(j) * sig.e(i) + rate.eg(j)
410 rc.e(j,i) - mid.e(i) + mid.eg + (zj(j) * (sig.e(i)A2 + sig.egA2)A0.5)
420 for t = 0 to 3
430 q(j,i,t) = bO(i) * (1 + rate.e(j,i)/100)A(t * 5)
440 c
-------
- 49 -
850 if i = 5 then print#2,"Methyl chloroform"
860 if i = 6 then print#2,"Halon 1301"
870 if i = 7 then print#2,"Halon 1211"
880 print#2,"Components of Joint Quantiles"
890 print#2," year";" c(.05)";" c(.25)";" c(.50)";
c(.75)";" c(.95)"
900 for t = 0 to 11
910 print#2,using"##### ",-year (t) ,q(l, i,t) ,q(2,i,t) ,q(3,i,t),
q(4,i,t),q(5,i,t)
920 next t
930 next i
940 for i = 1 to 7
950 printf2," "
960 if i = 1 then print#2,"CFC-11"
970 if i = 2 then print#2,"CFC-12"
980 if i = 3 then print#2,"Carbon tetrachloride"
990 if i = 4 then print#2,"CFC-113"
1000 if i - 5 then print#2,"Methyl chloroform"
1010 if i = 6 then print#2,"Halon 1301"
1020 if i = 7 then print#2,"Halon 1211"
1030 printf2,"Quantiles of Chemical Production"
1040 print#2," year";" c(.05)";" c(.25)";" c(.50)";
c(.75)";" c(.95)"
1050 for t = 0 to 11
1060 print#2,using"##### ";year(t),c(1,i,t),c(2,i,t),c(3,i,t),
c(4,i,t),c(5,i,t)
1070 next t
1080 next i
-------
- 51 -
REFERENCES
Hammitt, James K., Kathleen A. Wolf, Frank Camm, William E. Mooz,
Timothy H. Quinn, and Anil Bamezai, Product Uses and Market Trends for
Potential Ozone-Depleting Substances: 1985-2000, The Rand Corporation,
R-3386-EPA, May 1986.
Hirshleifer, Jack, Price Theory and Applications, Prentice-Hall, Inc.,
Englewood Cliffs, N.J., 1976
Kahneman, Daniel, Paul Slovic, and Amos Tversky, Judgment Under
Uncertainty: Heuristics and Biases, Cambridge University Press,
Cambridge, 1982.
Leontief, Wassily, "Population Growth and Economic Development:
Illustrative Projections," Population and Development Review, Vol. 5,
No. 1, 1979.
National Academy of Sciences, Halocarbons: Effects on Stratospheric
Ozone, Washington D.C., 1976.
National Academy of Sciences, Protection against Depletion of
Stratospheric Ozone, Washington B.C., 1979.
National Academy of Sciences, Causes and Effects of Stratospheric Ozone
Depletion: An Update, Washington D.C., 1982.
National Academy of Sciences, Causes and Effects of Changes in
Stratospheric Ozone: Update 1983, Washington B.C., 1984.
Palmer, Adele R., William E. Mooz, Timothy H. Quinn and Kathleen A.
Wolf, Economic Implications of Regulating Chlorofluorocarbon Emissions
from Nonaerosol Applications, The Rand Corporation, R-2524-EPA, June
1980.
Quinn, Timothy, Kathleen A. Wolf, William E. Mooz, James K. Hammitt,
Thomas W. Chesnutt, and Syam Sarma, Projected Use, Emissions and Banks
of Potential Ozone-Depleting Substances, The Rand Corporation,
N-2282-EPA, January 1986.
Ramanthan, V., R. J. Cicerone, H. B. Singh, and T. J. Kiehl, "Trace
Gases and Their Potential Role in Climatic Change," J. Geophysical
Res., Vol. 90, June 20, 1985.
U.S. Environmental Protection Agency, "Ozone-Depleting
Chlorofluorocarbons, Proposed Production Restriction," Federal
Register, Vol. 45, No. 196, October 7, 1980, pp. 66726-66734.
------- |