A RAND NOTE
          Prepared for
                                   SOCIAL COST OF TECHNICAL  CONTROL OPTIONS TO
                                   REDUCE EMISSIONS  OF POTENTIAL  OZONE DEPLETERS
                                   IN T!               ':   AN UPDATE
                                   F.  Camm,  T.  H.  Quinn,  A.  Bamezai,  J.  K. Hammitt,
                                   M.  Meltzer,  W.  E.  Mooz,  K.  A.  Wolf
                                   May 1986
                                   N-2440-EPA
The U.S. Environmental Protection Agency

-------
 The  research  in this Note is sponsored by  the U.S.  Environmental
 Protection Agency under  Cooperative Agreement No.  CR811991-02-0.
.The Rand Publications Series: The Report is the principal publication doc-
umenting and transmitting Rand's major research findings and final research
results.  The Rand Note reports  other outputs of sponsored research for
general  distribution. Publications of The Rand Corporation do not neces-
sarily reflect the opinions or policies of the sponsors  of Rand research.
                  Published by The Rand Corporation

-------
A RAND NOTE
          Prepared for
   Rand
       1700 MAIN STREET
        P.O. BOX 21M
   SANTA MONICA,CA 90406-2138
                                   SOCIAL  COST OF TECHNICAL CONTROL OPTIONS TO
                                   REDUCE  EMISSIONS OF POTENTIAL OZONE DEPLETERS
                                   IN THE  UNITED STATES:   AN UPDATE
                                   F.  Camm,  T.  H.  Quinn,  A.  Bamezai, J. K. Hammitt,
                                   M.  Meltzer,  W.  E.  Mooz, K. A. Wolf
                                   May 1986
                                    N-2440-EPA
The U.S. Environmental  Protection Agency

-------
                                - iii -
                               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 and J. K. Hammitt, An Analytic Method for Constructing
Scenarios from a Subjective Joint Probability Distribution,
N-2442-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.

-------
                                -  vii -

                               SUMMARY
     Photochemical models of the stratosphere suggest  that
chlorofluorocarbons and certain related chemicals  may  deplete
stratospheric ozone.  We call these chemicals "potential  ozone
depleters" or PODs.  Depletion of stratospheric ozone  may impair human
health, speed the degradation of certain man-made  materials,  reduce crop
yields, and have a broad range of negative ecological  effects.   Human
emissions of PODs may speed the depletion of stratospheric  ozone and
hence may contribute to negative effects associated with  depletion.
Human emissions of PODs can be reduced, however, only  by  reducing the
use of these chemicals and forgoing the benefits that  they provide.
     Policymakers in the United States and elsewhere must determine
whether the potential threat from human emissions  of PODs into the
atmosphere is large enough to warrant the costs of reducing their use.
This Note provides information relevant to that decision by quantifying
the social costs and effects of exercising specific technical options to
reduce the use of PODs in the United States.  It updates earlier Rand
work on this issue by updating the prices and cost of capital relevant
to these options and by incorporating new information from industry on
the cost and effectiveness of technical options.
     We rely on market prices to determine the social value of products
made with PODs and the social costs of the inputs used with PODs in
production processes or used to produce PODs themselves.  This is  a
common approach in social cost-benefit analysis and works so long  as
externalities are not important in the markets we examine.  Where
externalities are present, we identify them, but do not attempt to
monetize them.  Relying on market prices allows us to value the POD at
the price at which private firms would voluntarily switch production
processes or consumers would switch products.  This amount, less the
cost of producing the POD, represents the social  loss of regulations
that might induce these decisions to reduce  dependence on PODs.

-------
                                - viii -

     As earlier Rand work did, we emphasize technical options that would
be voluntarily adopted at "low to moderate" POD prices.  This focuses
the analysis on options that impose low to moderate social costs,
presumably the options that the United States should exercise first in
any attempt to reduce the use of these substances.  Specifically, we
consider only those options that might voluntarily be adopted at POD
prices below $5.00 a pound.  This price is five to 15 times the levels
of 1984 prices for the chemicals examined.
     A fairly small number of options account for most of the reductions
we examine.  Recovery and recycling of PODs in the manufacturing
processes in which they are used as inputs reduces requirements for
these chemicals and reduces their emissions into the atmosphere.
Manufacturers typically weigh the costs of new investment and operating
costs against the savings that recovery and recycle of PODs allows.  Use
of equipment and methods designed to reduce emissions directly involves
a similar tradeoff.  Substitution of alternative materials for the PODs
requires a comparison of their costs and their effects on the
productivity of a manufacturing process.  This last option may involve
social costs that our methodology does not capture because it often
involves substitution away from PODs to chemicals that may be dangerous
for other reasons.   Among the alternatives to PODs are pentane,
methylene chloride, and other chlorinated solvents that pose fire and
health hazards and disposal problems that may not be reflected in their
market prices.  This is the only place where externalities affect our
analysis.
     The technical options that our quantitative analysis suggests that
firms would adopt voluntarily at POD prices below $5.00 a pound would
have the following effects:

     •   Cut the use of CFG-11 in slabstock foam manufacturing by about
         37 to 63 percent at prices below about $2.00 a pound, a
         fourfold increase in price;
     •   Phase out the use of CFG-12 in the manufacture of thermoformed
         polystyrene sheet at prices below $1.00 a pound, a 50 percent
         increase in price;

-------
                                 - ix -

     •   Cut the use of CFC-12 in retail food refrigeration by about
         half at prices below about $1.20 a pound, an 80 percent
         increase in price;
     •   Reduce some use of CFC-12 in other refrigeration and air
         conditioning applications at prices below $3.65 a pound, about
         a 450 percent increase in price; and
     •   Cut the use of CFC-113 in cleaning and drying applications by
         over 90 percent at prices below $5.00 a pound, about a 460
         percent increase in price.

We also believe, on the basis of less quantitative analysis, that the
use of CFC-11 in molded foam manufacturing could be phased out, that
currently exempt uses of CFC-11 and -12 in aerosol applications could be
cut in half, that use of CFC-12 in sterilants could be cut in half, that
better servicing practices could reduce the emissions of CFC-11 and -12
during servicing of certain refrigeration and air conditioning
equipment, and that use of CFC-12 in liquid food freezing could be
phased out.
     Taken together, these suggest that technical options with low to
moderate social costs would be more effective at reducing the use of
CFC-113 than the use of CFC-11 and CFC-12.  That is true because few
good substitutes exist for CFC-11 and -12 in the majority of their
applications.  As a result, the social cost of reducing the use of these
chemicals will be quite high in most applications.  The technical
options considered in this Note allow a total reduction of about 6 to  16
percent in the use of CFC-11, 6 to 35 percent in CFC-12, and 75 to 80
percent in CFC-113 in the United States.

-------
                                 - xi -
                         ACKNOWLEDGMENTS
     John Hoffman and Stephen Seidel helped us  frame the analysis  in
this Note.  Jan Acton helped facilitate its production and review  under
tight deadlines.  David Rubenson helped collect industry data  on
refrigeration applications.   Hugh Farber and Leland Johnson provided
detailed comments on an earlier draft.   The Alliance for Responsible  CFG
Policy and W. J. Rhodes also provided helpful comments.   Mary  Vaiana
helped prepare the presentation of this Note to the Environmental
Protection Agency's March 1986 workshop, "Protecting the Ozone Layer."
Participants in that workshop provided helpful  feedback.  Alyce Shigg
oversaw production of the many drafts underlying this Note and Patricia
Bedrosian edited the final draft.  More generally, much of the detail
provided here is based on interviews with many knowledgeable industry
officials who were generous with their time.  We thank them all and
retain responsibility for any errors that remain.

-------
                                -  xiii -
                              CONTENTS
PREFACE  	  iii

SUMMARY  	  vii

ACKNOWLEDGMENTS  	   xi

FIGURES  	   xv

TABLES  	 xvii



Section

   I.   INTRODUCTION  	     1


  II.   BACKGROUND AND METHODS  	     4

       Social Cost and Demand for PODs	     4
       Implications for Measuring a Demand Schedule  	     8
       Prices at Which Users Voluntarily Exploit Technical
        Options  	    13
       Summary  	    15

 III.   FLEXIBLE FOAM  	    16

       Alternative Blowing Agents  	    16
       Recovery and Recycle  	    17
       Quantitative Analysis of Technical Control Options  	    17

  IV.   RIGID FOAM  	    20

       Alternative Blowing Agents  	    20
       Recovery and Recycle During Manufacture  	    21
       Product Substitutes  	    21
       Quantitative Analysis of Technical Control Options  	    21

   V.   SOLVENTS  	    .74

       Vapor Recovery  	    24
       Recovery from Waste  	    25
       Product Substitutes  	    26
       Quantitative Analysis of Technical Control Options  	    27

-------
                                - xiv -
  VI.  REFRIGERATION AND AIR CONDITIONING  	   33

       Options Documented with Quantitative Analysis  	   33
       Other Options To Reduce POD Emissions  	   34
       Quantitative Analysis of Technical Control Options  	   36

 VII.  MISCELLANEOUS APPLICATIONS  	   39

       Exempt Aerosol Propellant Uses  	   39
       Liquid Food Freezing  	   41
       Sterilants  	   41
       Halon 1301  	   42

VIII.  CONCLUSIONS AND DIRECTIONS FOR FUTURE WORK  	   43
Appendix

   A.  POTENTIAL PROBLEMS IN COMPARING ALTERNATIVE
       POLICIES TO REDUCE POD USE 	   53
   B.  MODELS AND DATA USED TO QUANTIFY TECHNICAL CONTROL
       OPTIONS 	   59
   C.  DIFFICULTIES IN USING U.S. RESULTS TO LOOK ABROAD 	   70

REFERENCES  	   75

-------
                                 -  XV  -
                               FIGURES
2.1.  Demand and willingness to pay  	    7




2.2.  Net social cost of a restriction on  POD use   	    8




2.3.  Engineering measure of the social cost of  reducing  POD use   .    10




2.4.  How consumer response affects the cost of  reducing  POD use   .    11




8.1.  Demand schedules for U.S. use of CFG-11   	    45




8.2.  Demand schedules for U.S. use of CFC-12   	    47




8.3.  Demand schedules for U.S. use of CFC-113   	    50

-------
                                -  xvii -
                                TABLES
3.1.  Technical Options To Reduce the Use of CFC-11  in  the
      Manufacture of Flexible Slabstock Foam  	    18

4.1.  Technical Options To Reduce the Use of CFC-12  in  the
      Manufacture of Thermoformed Polystyrene Sheet   	    22

5.1.  Where Losses Currently Occur in the U.S.  Market for
      CFC-113 	    27

5.2.  Technical Options To Reduce the Use of CFC-113 in Cleaning
      and Drying Applications  	    29

6.1.  Technical Options To Reduce the Use of CFC-12  in  Air
      Conditioning and Refrigeration Applications  	    36

B.I.  Data Inputs for Flexible Slabstock Foam  	    61

B.2.  Data Inputs for Thermoformed Polystyrene Sheet  	    64

B.3.  Data Inputs of CFC-113  	    68

-------
                                 - 1 -
                          I.   INTRODUCTION
     Over the last 12 years, photochemical models of the atmosphere have
suggested that global human emissions of certain chlorofluorocarbons
(CFCs) and related chemicals may deplete stratospheric ozone.   We will
refer to these chemicals as potential ozone depleters or PODs.   We focus
our attention here on CFC-11, -12, and -113, but we also consider
technical options relevant to methyl chloroform and Halon 1301.
     A sufficient reduction in stratospheric ozone could have  a wide
variety of detrimental effects, including significant threats  to human
health.1  However, scientists have not reached a consensus that these
effects do in fact occur or even that they are likely in the future.
Nonetheless, if the effects do occur, they could be serious enough to
warrant countermeasures that would reduce human emissions of these PODs
in the absence of any scientific consensus.  It would be easier to
justify such countermeasures politically if they did not impose too
serious a social cost.  In fact, a number of governments have already
taken limited steps to reduce current emissions of certain CFCs on
precisely these grounds.2  This Note updates earlier Rand Corporation
estimates of the costs of undertaking a variety of additional control
measures in the United States.*
     A cost is imposed when a control measure forces an individual or
company to change its behavior.  That cost may result when a product
     'For details, see National Academy of Sciences (1976, 1979, 1982,
1984) or Raraanthan et al. (1985).
     2Bans on the use of CFCs in most aerosol products are now in effect
in Canada, Norway, Sweden, and the United States.  The European Economic
Community agreed that, by the end of 1981, CFC aerosol use would be
reduced by 30 percent relative to 1976 levels.  Australia has reduced
CFC aerosol use through voluntary agreements with industry.  West
Germany and the Netherlands have labeling programs in effect to
encourage the use of alternative propellents.  Japanese industry has
reportedly agreed not to expand the use of CFCs as aerosol propellants.
     3See especially Palmer et al. (1980), Palmer and Quinn (1981), and
Mooz et al. (1982).

-------
                                  - 2  -
 previously produced with PODs must be produced in a different way.
 Because  the  alternative was not used before, it presumably will cost
 more;  this difference  in production cost is what interests us.  Of
 course,  if a product costs more to produce without PODs, its price will
 rise,  encouraging the  consumers of that product to seek less expensive
 alternatives.   If consumers can find alternatives, they limit the cost
 imposed  by a control measure.  The mitigating effect of product
 substitution also interests us.  Substitution away from PODs in
 production or consumption may increase the use of substances like
 methylene chloride or  pentane that are themselves dangerous or
 potentially  dangerous.  Because new users may not properly recognize the
 full extent  of  their dangers, the private cost to companies or
 individuals  of  using these substances may underestimate their total cost
 to society.h  We identify why our cost estimates might fall short of the
 full costs relevant to society but we do not attempt the difficult and
 controversial task of  quantifying the extent of the shortfall.
     The Rand Corporation's previous work focused on identifying control
 options with "small to moderate" social costs in the United States.
 These  are presumably the first options that should be employed if U.S.
 policymakers decide to reduce the emissions of PODs.  This update has a
 similar  focus.  Where possible, we use engineering cost data to estimate
 the effects  of  control measures on production costs.  We could not
 obtain such  information for some options that appear to impose only
 modest costs; we identify these options and recommend future work to
 identify their  costs more clearly.
     The estimates offered here improve on Rand's previous estimates by
 reflecting recent changes in economics and technology.  The Rand
 Corporation's earlier  estimates of costs date variously from 1976, 1980,
 and 1982.  At the very least, nominal and real prices of equipment,
 labor, energy,  and, of course, individual chemicals have changed since
 those estimates were made.*  Recent and prospective changes in the tax
     *An economist would say that their use imposes an "externality"
because it imposes a cost on people who are not a party to the decision
to use them.  The monetized value of this externality captures the
additional cost not examined here.
     8Bureau of Labor Statistics data suggest that equipment costs in
the chemical industry rose 80 percent from 1976 to 1984 and 31 percent
from 1980 to 1984.  Labor costs in the chemical industry rose 88 percent

-------
                                 - 3 -
law and in real interest rates have changed the cost of capital  to
companies that must make new investments to reduce their use of  PODs.6
Some options considered in earlier work have been adopted and foreclose
the possibility of further reductions.   And additional discussions with
industry have refined Rand's understanding of the remaining available
options and the costs associated with them.  This update incorporates
new information on prices, cost of capital, and industry options.
     Section II provides background on a variety of issues relevant to
the specification and estimation of the social costs of control  options.
Sections III through VII briefly review the control options for  flexible
foam, rigid foam, solvents, refrigeration and air conditioning,  and
other applications that appear to have low to moderate costs in  the
United States.  They update Rand's information on their costs and
effectiveness in reducing emissions from PODs.  Section VIII summarizes
our results on the extent of reductions and their costs for the United
States and suggests directions for future work.
     Appendix A briefly reviews some methodological issues that arise
when results like those developed here are used to compare specific
policy alternatives.  It emphasizes the specific character of the
results reported here and the need for additional work before these
kinds of results are used to compare different kinds of policies.
Appendix B reviews briefly the methodology and assumptions used to
develop results for foam manufacturing and solvent use.  Appendix C
briefly reviews issues relevant to the use of these U.S. data to  infer
the costs and effects of applying similar  control options  in other parts
of the world.
from 1976 to 1984 and 33 percent from  1980 to  1984.  Energy costs rose
147 percent from 1976 to 1984 and  14 percent from  1980 to  1984.  CFC-11
rose in price 50 percent from 1976 to  1984; CFC-12 and -113 prices rose
63 and 44 percent over the same period.
     'The real pretax cost of capital  normally used in industry in 1984
was about 15 percent, 25 percent lower than the  level assumed  in Palmer
et al. (1980).

-------
                                 - 4 -
                  II.   BACKGROUND AND METHODS
     We use the economic tools of social cost-benefit analysis  to
measure the minimum social cost of reducing the production and  use  of
specific PODs.  In particular, we rely on the close relationship  between
demand and supply schedules for PODs and the net social costs of
eliminating activities associated with those schedules.  We assume  that
activities relevant to any particular chemical are eliminated in
ascending order of their social cost to assure that the use of  this
chemical cannot be reduced in any lower cost way.1  This section  briefly
reviews our approach to social cost-benefit analysis and explains how we
use demand and supply schedules to implement that approach. It then
examines how we quantify points on the demand curve for each chemical.

SOCIAL COST AND DEMAND FOR PODS
     Benefit-cost analysis in market economies normally uses market
prices to measure the social costs of inputs and the social value of
outputs.  Prices indicate how much someone would have been willing  to
pay for the inputs in another use.  They also reflect consumers'
willingness to pay for the outputs.   Because benefit-cost analysis  is
ultimately based on individuals' willingness to pay for goods given the
current distribution of income, free market prices normally provide the
proper information for such analysis.  And we will lean heavily on
market prices wherever possible.2  The most important circumstance  that
     1This is a restrictive assumption.   The types of environmental
policies most often observed in the United States and elsewhere do not
pursue such priorities.  Hence, our analysis offers a lower bound on the
social cost likely to be associated with any actual government policy.
Future Rand analysis will examine alternative ways to reduce the use and
emissions of PODs and compare their social costs.  The estimates here
can best be thought of as benchmarks for more complex policies.  For a
further discussion of this and related issues, see Appendix A.
     2This paragraph is obviously a broad statement about complex
issues.  The most direct and useful defense of this approach is
Harberger (1971).  This reference also discusses important caveats in
using the approach.

-------
                                 - 5 -
might raise doubts about this approach is one in which an input  or
output affects people who either are unaware that they are being
affected or who—for whatever reason—can do nothing about the effects.
We believe that this is most likely to present a problem if a
restriction on the use of PODs encourages consumers or producers to
increase their demand for certain other chemicals whose true effects  are
not fully known to the workers who work with them or to the consumers
who use products made from them.  For example, methylene chloride causes
cancer in animals and may be a human carcinogen.  If restrictions on  the
use of CFC-11 encourage greater use of methylene chloride, that  use may
impose a greater cost on workers using it than they realize and  hence
may lead them to underprice their labor services in working with
methylene chloride.  That is, if workers underprice their services, the
true social cost of using methylene chloride in production is higher
than the cost we would measure using market prices.  Asking how  much
higher opens a controversy that cannot be resolved here.  Ve avoid this
by relying on market prices and simply noting where reasons may  exist to
expect divergences between these prices and the true social values
relevant to our analysis.
     A decision to rely on market prices for cost information simplifies
our problem in an important way.  Because the measures of cost that we
use are the same as the measures individuals use in making decisions, it
allows us to infer social costs from people's behavior.  In particular,
it allows us to use the fact that the demand schedule for a particular
POD reflects the willingness to pay for that substance.  Similarly,
supply schedules reflect willingness to pay to use the substance—or the
resources used to produce it—in alternative uses.  When production and
use of a POD is restricted, the amount that users would have been
willing to pay for the restricted quantity is a measure of how much this
restriction hurts them.  Users of the POD in other applications or of
the resources used to produce it would be willing to pay some amount to
obtain the restriction to increase the supplies available for their own
use; this is a measure of how much they would benefit from a
restriction.  The net social effect—neglecting of course the value of
effects on stratospheric ozone--is the difference between these two
quantities.  And this difference can be expressed  in  terms of areas

-------
                                 - 6 -
under the segments of the demand and supply schedules affected by the
particular restriction in question.
     To see this, consider the relationship between willingness to pay
and areas under demand curves.  The total demand for a POD is really an
aggregation of individual demands for its use in particular
applications.  For example, the demand for CFC-11 is an aggregation of
demands for specific uses in foam blowing, aerosols, and so on.
Further, the demand for foam blowing is really an aggregation of demands
for foam blowing in different plants with different production
technologies and products.  For each specific demand, there is a price
of CFC-11 at which that demand stops.  At a high enough price, foam
blowers switch to alternative blowing agents or final consumers respond
to the high product prices forced by the higher CFC-11 price and spend
their money on other products.  The price at which use ends is the
maximum price that foam blowers and final consumers are willing to pay
for the POD in a particular use.
     This "willingness to pay" varies according to each use.  Imagine a
rectangle associated with each specific use, with the "switch price" on
the vertical dimension and the level of demand in the use on the
horizontal dimension.  The area of this rectangle is what POD users and
final consumers would be willing to pay to avoid losing this use of a
POD.  And the sum of many analogous areas is the total willingness to
pay to avoid losing all the relevant uses of the POD.
     In the end, a demand schedule can be thought of in terms of a
collection of these rectangles.  (See Fig. 2.1.).  Place the tallest
rectangle to the left and add progressively shorter rectangles; the
aggregate demand schedule for a POD (or any other substance) is simply
the upper boundary of this collection of rectangles.  As a result, we
can measure the willingness to pay to avoid a restriction on using a POD
by looking at the area under the sections of that demand schedule that
cannot be satisfied when use is restricted;
     By similar reasoning, it can be shown that the area under a supply
schedule measures the willingness to pay for the POD or its inputs in
other uses.   For simplicity, we will always assume here that the supply
schedule for PODs is flat so that the segments of the supply schedule
relevant to a restiction are not an issue.  This simple assumption

-------
                                 -  7  -
                              Demand schedule
                            Level of demand
               Fig. 2.1 -- Demand and willingness to pay
allows us to measure the social costs of a restriction as the areas
between the segments of the demand schedule affected and the supply
schedule.  For example, if a restriction eliminated the uses of a POD
associated with segments A and B of the demand schedule in Fig. 2.2,
areas AA and BB would measure the social cost of this restriction.
Because we assume here that restrictions are imposed in a least cost
manner, our measures of social cost will always involve triangular areas
bounded by the demand and supply curve and the restriction with the
highest incremental social cost.

-------
                                 - 8 -
                           Level of supply, demand
        Fig. 2.2 -- Net social  cost  of  a  restriction on POD use
IMPLICATIONS FOR MEASURING  A DEMAND SCHEDULE
     Normally, economists  develop empirical demand schedules by using
historical data and econometric techniques to estimate them.  But the
available data are not  good enough to use formal econometric methods to
estimate reliable demand schedules,  even at the aggregate level and for
the major PODs--CFC-ll  and -12.   Even if they were, relevant prices have
not varied enough recently to  measure a demand schedule over the range
of prices that concerns us.  Hence,  we adopt an alternative approach.

-------
     The approach is easiest to understand in terms of incremental
reductions in demand from the existing level of demand without
regulations.  That is, we estimate^ ^rectangles like those discussed above
more interms of their contribution to reducing existing demand than in
terms_of their contribution to aggregating up from a zero base.  Three
steps are potentially important to estimating the costs of these
reductions.
     First, we should consider specific control measures that might be
taken to reduce demand, like recovering and recycling CFC-11 in foam
blowing in large plants.  For these, we start by assuming that reduction
is relevant only at the production level; when a control measure raises
production costs, thereby forcing a price increase, consumers do not
react by cutting consumption.  For each control measure, we use
engineering cost data and market data to estimate the vertical (cost)
and horizontal (size of reduction) components of each rectangle.  Ranked
from lowest to highest cost per pound, these rectangles look like those
in Fig. 2.3.  To minimize the social cost of achieving a given reduction
in the use of this POD, it would be most appropriate to undertake option
A first, option B next, and so on.  Adding up the areas of these
rectangles provides a measure of willingness to pay; subtracting the
relevant area under the supply curve yields a measure of social cost.
But these measures of willingness to pay and social cost could be too
high unless we took the second step.
     In the second step, we recognize that in fact consumers do respond
to higher prices.  Perhaps the easiest way to think about this is to
realize that each rectangle in Fig. 2.3 is in fact a composite of many,
each representing the reduction in POD use for a particular consumer
class within this rectangle when the control option in question is
implemented.  For some consumers, price increases in the POD below the
vertical height of the rectangle—the switch price assuming no consumer
response—are sufficient to force price rises that encourage consumers
to drop their consumption of these products in favor of other products.
For these consumers, the switch prices relevant to them, and the
willingness to pay to avoid an imposed reduction  in POD use, are  lower
than the switch price for the large rectangle.  Actual willingness to

-------
                                  -  10  -
                                     B
                                                    Demand
                                                    schedule
                                                    (inverted)
                                                      Supply
                                                      schedule
                              Level of reduction
 Fig. 2.3 -- Engineering measure of the  social  cost  of  reducing POD use
pay to avoid the restriction  is measured  by  the ordered rectangles in
Fig. 2.4(a).  If we had stopped with  the  first  step,  we would have
overestimated the relevant willingness  to pay by the  shaded areas.
     Of course, there is no particular  reason to rank the consumer-
related rectangles of Step 2  only within  the larger rectangles.   A
complete ranking yields the arrangment  in Fig.  2.4(b).   The difference
between panels (a) and (b) reflects the fact that to  minimize the social
cost of achieving a given reduction in  POD use, it may be appropriate to

-------
                                     - 11  -
               [a]
            8
            d
                                  Level of reduction
                                                 Demand
                                                 schedule
                                                 (inverted)
                                                     Supply schedule
                                  Level of reduction
Fig.  2.4  --  How  cor^sttrtier response affects  the cost of reducing  POD use

-------
                                 - 12 -
eliminate consumption by some groups affected by option B, the second
best option, before eliminating all consumption relevant to the "best"
option, option A.  That is, simply using mandatory controls to implement
one option after another misses cost savings that could be achieved if
consumers with lower switch prices could be identified and dealt with
before moving on to consumers with higher switch prices.
     A third step would recognize that our information about control
options and consumer response is likely to be incomplete.  That is,
there may be a control option, like option C in Fig. 2.3, that escapes
current notice because its effect on demand is so small or because it
fills a niche that takes time to discover and exploit.  Concentrating
only on the "significant" options we now know about ensures that we
overestimate the ultimate willingness to pay to avoid a cut in POD use.
     In the end, the ranking shown in Fig. 2.4(b) is just a demand
schedule, with quantity reductions instead of total quantity on the
abscissa.  It allows us to walk up the demand schedule from the point of
current consumption instead of walking down it from the point of zero
consumption; these alternatives differ only in perspective.  In effect,
the rectangles in Fig. 2.4(b) tell us the true character of the
rectangles that a demand schedule bounds.
     In the sections that follow, we focus on the first step.  Among the
reduction options we have been able to identify, it appears that the
second step—reflecting the response of final consumers to hi her
prices — is likely to make a difference for only one POD use--the use of
CFC-12 in extruded polystyrene she t.  And the difference is likely to
be small and to affect only a portion of the demand schedule for CFC-12.
The third step is obviously more difficult to quantify.  It is likely to
become more important as time passes, allowing time for innovation of
new options.  It is also more likely to be important in the evaluation
of price-oriented policies than in that of mandatory controls.1  The
analysis that follows takes no account of this third step.  As a result,
we can expect the minimum social cost of reducing the use and emissions
of the PODs considered here to fall below our estimates as time passes.
     3For a discussion of this issue, see Appendix A.

-------
                                 -  13  -
PRICES AT WHICH USERS  VOLUNTARILY EXPLOIT TECHNICAL OPTIONS
     Our analysis focuses on the  amount  of PODs associated with specific
production processes  and the prices  of PODs  at which users voluntarily
change production processes.  These  prices are the switch prices
referred to above.4
     The amount of a  POD that a producer must use to produce any
particular quantity of a product  is  not  fixed.  This is true even if the
final product must contain  a certain amount  of the POD.  Certain amounts
of a POD are lost in  the manufacturing process/ through emission during
manufacture, through  the disposal of contaminated waste not embodied in
the final product, or through the disposal of scrap product whose
quality is too low to be sold. Through  greater care or additional
investment, these kinds of  losses can be reduced.  But greater care and
investment are costly.  Users of  PODs must weigh the cost of reducing
their use of PODs against the cost of acquiring PODs.  We compare the
costs of alternative  ways of using PODs  in manufacturing and services
and determine the prices at which it pays users to switch from one
method to another.
     We typically start by  looking at operations in a plant of some
particular size.  We  then assume  that the level of production in the
plant is fixed; demand for  final  product is  insensitive to changes in
production method.*  We then compare the full  annualized costs of
achieving the assumed production  level  using different-production
methods.  That is, where new investment is  required to  implement a
particular method, we calculate the  return to  capital  required each year
to cover the .cost of the investment  and include  that  as the cost of
investment.  Operating and  fixed annual charges  are dealt with directly.
     "The approach draws heavily on earlier Rand analysis reported in
Palmer et al. (1980), Palmer and Quinn (1981), and Mooz et al.  (1982).
This subsection briefly summarizes discussions offered in much greater
depth in these earlier documents.  Appendix B provides additional
detail.
     *We review the possibility that a change in process could change
costs, product prices, and hence demand; this is important in only one
case.

-------
                                 - 14 -
     The price of the POD used in production is the only cost of direct
interest to us.  With some reasonable simplifying assumptions, we can
express the annualized cost of any production method as a simple linear
function of that price.  That is, C. = a. + b.P, where C. is the
annualized cost for the ith production method, P is the per unit cost of
the POD in question, and a. and b. are constants.  At any price, P, the
user obviously picks the production method with the lowest C..  As the
price rises, a user moves progressively toward methods with higher
ratios of a. to b..  We capture the price at which a user switches from
one method (i) to another (j) by calculating the price, P*(i,j), at
which the costs for the two methods are equal:
                     P*(i,j) = (a.. - a^/Cb. - bi)
If no other method (say, k) is less costly than these at this price,
this switch price is relevant to us and important to the construction of
demand schedules as discussed above.6
     As price rises, we can expect—and in fact observe—a series of
prices at which the use of a POD within an industry falls.  This is true
even if the industry has only one alternative to current practice.  Many
switch prices may be relevant within the industry because switching
often occurs at different kinds of plants at different prices.
Differences in scale or product mix can lead to different costs and
hence different a.^ and b. for the production methods used in different
plants.  The reduction in POD use in this case simply represents
changing practice in one part of the industry in question.7  Of course,
any particular plant may move through several forms of conservative
     *If another method was less costly, then this switch point would be
irrelevant because the methods relevant to the switch are dominated by
another method at this price.  We seek only the lowest cost alternatives
and POD prices at which their costs are equal.
     7We might reasonably expect changes in the price of a POD to affect
the competitiveness of different kinds of plants.  Plants that can most
easily adjust to rising POD prices should presumably be able to compete
more effectively against other plants in the same industry.  We make no
attempt to take account of shifts in market share that might result from
this kind of cost competition.  That makes our analysis conservative.

-------
                                 - 15 -
practice as the POD price rises.   Where this is possible,  it  simply
expands the number of prices at which POD use falls.
     In a few cases, our analysis suggests that certain practices  not  in
universal use today should already be in use; they have costs that
appear to be lower than the costs of current practice.   In some of these
cases, representatives from private industry have concurred with this
analysis, but would like to see others move first to control  the risks
associated with innovation.  In these cases, we generally cannot
determine exactly how much of the industry has-moved toward what appears
to be a cost effective innovation.  Hence we bracket the correct share
that has innovated by using alternative base cases.  In one,  all cost
effective innovations have been made; in the other, none have.

SUMMARY
     We use standard methods of social cost-benefit analysis  to examine
the net social cost of reducing the use of PODs in specific ways.   Our
analysis focuses on identifying the least cost way to do this for each
chemical and measuring the cost of the corresponding reductions.
Because we rely on market prices to define values relevant to cost-
benefit analysis, our analysis is consistent with a behavioral view of
these markets; that is, our analysis yields demand schedules that show
what actions users of PODs would take if the prices of these PODs were
to rise.  Future work will examine the costs associated with alternative
methods for reducing the use and emissions of PODs.

-------
                                 - 16 -

                         III.   FLEXIBLE FOAM
     CFC-11 is used as an auxiliary blowing agent in flexible urethane
 foam for the manufacture of molded foam and slabstock products.   Two
 technical options exist for reducing CFC emissions from these processes:
 substituting auxiliary blowing agents, and recovering and recycling the
 emitted CFC.  This section reviews these options briefly and presents
 quantitative data on their effects on CFC-11 use and the prices  of
 CFC-11 at which manufacturers would adopt them voluntarily.

 ALTERNATIVE BLOWING AGENTS
     Methylene chloride can be substituted for CFC-11 as a blowing agent
 for many slabstock foams.   The materials costs of methylene chloride
 formulations are currently lower than those of CFCs for many grades of
 slabstocks.  Quality control of soft foams blown with methylene  chloride
 is more difficult, however, sometimes leading to higher scrap rates.
 This is especially the case in smaller manufacturing plants that often
 do not have the technical expertise for using the chemical efficiently.
 In addition, methylene chloride is presently under regulatory scrutiny
 and its use might be restricted in the future.
     From a technical standpoint, it would be difficult to achieve
 complete replacement of CFC by methylene chloride unless consumers were
 willing to accept lower quality foam.  Reasons for this range from the
 unavailability of methylene-chloride-based formulations for some
 specialty foams to the inability of some foamers to handle the technical
 problems of blowing agent conversion.  Industry observers indicate that
 perhaps 75 percent of flexible foam blowing could be achieved with
 methylene chloride.
     Until recently, industry sources suggested that substitution among
blowing agents was more difficult in molded foams.  The quality  of foam
blown with alternative agents tended to be poor and perhaps
unacceptable.   New information suggests that substitutes are available
that can be implemented without much cost penalty.  The specific nature
of these substitutes remains confidential.  We do not attempt to

-------
                                 -  17 -
quantify the opportunities  for  substitution here beyond assuming that
the use of CFC-11 would be  eliminated in molded foam blowing if the
price of CFC-11 rose to the top of the range we consider, $5.00 a pound;
we cannot say at what price the switch would actually occur.

RECOVERY AND RECYCLE
     Flexible foams emit essentially all of their auxiliary blowing
agents before leaving the manufacturing plant.  Collecting and reusing
these emissions would prevent them from being vented into the
atmosphere.  Flexible slabstock lines appear particularly suited for
such collection, since the  foam is manufactured in a long tunnel
equipped with ventilation fans  that collect exhaust gases and discharge
them outside of the plant.   Measurements by DuPont indicate that in a
well designed, modern slabstock plant, CFC collection efficiencies using
these ventilation systems may already be between 33 percent and 53
percent.l
     Once collected, CFC may be recovered by carbon adsorption.  In this
process, CFC-laden air is passed over and adsorbed by beds of carbon.
The remaining air is exhausted, and CFC is desorbed from the carbon
using steam.  This technology is presently successfully and economically
practiced in a number of nonfoara industries.

QUANTITATIVE  ANALYSIS OF  TECHNICAL CONTROL  OPTIONS
     Our quantitative analysis  of options to reduce the use of CFC-11 in
flexible slabstock foam focuses on three different sized plants, which
would consume respectively  150,000, 225,000, and  1,200,000 pounds  of
CFC-11 annually if only CFC-11  were employed as an auxilliary blowing
agent.  To reflect information  from  industry sources,  however, we  assume
that each of these plants presently uses 50 percent methylene chloride.
Each plant has the option of increasing its use of methylene chloride to
75 percent and/or installing recovery and recycling equipment for  CFC-11
and methylene chloride.  Despite differences  in the sizes of the plants,
industry sources indicate that  this  equipment  is  equally costly  for  all
plants, giving large plants a significant advantage  in adopting  this
      'See Palmer et al. (1980), p. 52.

-------
                                 - 18 -
option.  We reflect that here.  Wherever it is used, this equipment
recovers half of the blowing agent emitted.  Flexible slabstock foam
accounts for about 14 percent of the U.S. market for CFC-11.2
     Table 3.1 displays the basic results of our analysis.  Each row
presents information on a particular action to reduce the use of CFC-11.
The first column shows the price at which users would take actions
voluntarily; this is of course also their willingness to pay for the
CFC-11 eliminated by each action.  The second column shows the size of
plant taking the action.  The third indicates what the action is; a
plant either "converts" to methylene chloride or "recovers" and recycles
CFC-11.  The fourth column indicates the share of the slabstock market
affected by the action.  The fifth shows the reduction in CFC-11 within
each sector achieved by the action.  The last column, calculated as the
product of- the numbers in the fourth and fifth, presents the percentage
reduction in the total slabstock market.  Multiplying the numbers in the
final column by 0.14 would yield the effects of these actions on total
U.S. use of CFC-11.

                               Table 3.1
          TECHNICAL OPTIONS TO REDUCE THE USE OF CFC-11 IN THE
                 MANUFACTURE OF FLEXIBLE SLABSTOCK FOAM
Price
(S/lb)
0.30
0.68
1.41
2.05
[a]
Plant
Size
Large
Medium
Large
Small
See text.
Action
Taken[ a]
Convert
Convert
Recover
Convert

% of
Slabstock
Market
53
28
. 53
19

» Reduction
in Plant Use
50
50
25
50

Total %
Reduction
26.5
14.0
13.3
9.5

     2Our analysis draws heavily on methods and data from Palmer et al.
(1980), pp. 53-58 and 267-270.  For a summary of our assumptions, see
this source and Appendix B.

-------
                                 - 19 -
     Note first that the current price of CFC-11 is $0.51 per pound,
suggesting that large plants should already be converting to higher use
of methylene chloride.  In fact, we observe this in many large plants.
We do not know exactly what share has converted today; we can resolve
this by considering two alternative base cases, one without any
conversion and another with total conversion among large plants.  We
take this approach in Sec. VIII.
     At somewhat higher prices, medium and then small plants voluntarily
convert.  Large plants also begin to recover and recycle their blowing
agents.  Other plants require much higher prices to do this voluntarily
because the economies of scale in recovery are substantial.
     In the end, all plants consider increasing their use of methylene
chloride before recovering blowing agents.  Only large plants use
recovery in significant numbers at low prices.  Within the price range
considered, conversion and recovery allow a reduction in the U.S.
slabstock foam industry's use of CFC-11 by about 37 to 63 percent and a
reduction of total U.S. use of CFC-11 of about 5 to 9 percent.

-------
                                 - 20 -
                           IV.   RIGID FOAM
      CFC-12  is used in rigid foam products, like polystyrene (PS) sheet
 and  board  and polyolefins, which are used extensively for packaging or
 temporary  containment of foods and other items.1  Three options are
 available  to reduce the use of CFC-12 in these applications:  alternative
 blowing  agents, CFC recovery and recycle during manufacture, and product
 substitution.  This section reviews these options briefly and then
 presents quantitative estimates of the cost and effectiveness of the
 first two  options.  Empirical information is less complete for the last
 option.

 ALTERNATIVE BLOWING AGENTS
     Extruded PS sheet is the primary candidate for using a substitute
 blowing  agent.  Pentane is currently being employed in a number of
 situations,  and available evidence suggests that it could serve as a
 blowing  agent in virtually all thermoformed sheet products.
 Thermoformed sheet products accounted for 81 percent of total PS sheet
 output in  1977, and 74 percent of CFC use in PS sheet production (Palmer
 et al.,  1980, p. 97).
     Use of  pentane has some drawbacks.   First,  conversion to pentane
 would require investment in new equipment.   Second, pentane blowing
 agents can pose a serious fire hazard to production workers, especially
 if the polystyrene resin is ignited.   Third, pentane has in recent years
 been the subject of several local regulatory actions, as a result of
 suspicions that as a low-boiling gasoline fraction, it contributes to
 the formation of photochemical smog.
     In spite of these drawbacks, industry observers and the analysis
 below suggest that the option of substituting pentane for CFC-12 in the
production of extruded PS sheet offers a good possibility of bringing
about a large reduction in CFC use.   Economically competitive facilities
     1CFC-11 is the most common POD used to blow rigid foam.   We do not
have information on any low-cost methods to reduce CFC-11 in this
application.

-------
                                - 21 -
for manufacturing and  marketing pentane-blown PS sheet already exist  for
virtually all applications  in which PS sheet is now used.

RECOVERY AND RECYCLE  DURING MANUFACTURE
     Largely in response  to potential regulations on pentane use,
recovery of manufacturing emissions is being investigated at this  time
by producers of both pentane and CFC-blown PS sheet.  Recovery
techniques are already used in the manufacture of extruded polypropylene
foam.  This process employs carbon adsorption technology, achieves an
overall recovery efficiency of 80 percent, and is economical at current
CFC prices (DuPont, 1978).  Although recovery from extruded PS sheet  is
economically less attractive, the large quantities of CFC-12 available
for collection in a single  plant and the probable absence of chemical
contaminants in the air stream suggest that recovery is possible as a
voluntary industry response to regulatory stimulus, and might be an
enforceable control candidate.

PRODUCT SUBSTITUTES
     Product substitutes  such as wood fiber products exist for foam
packaging and container applications, and are economically competitive
in a limited group of  those applications.  In uses  such  as egg cartons
and food service packages and trays,  for instance,  foam  and wood fiber
products sometimes compete  side by side  in the same supermarket.

QUANTITATIVE  ANALYSIS OF TECHNICAL CONTROL OPTIONS
     Like the flexible urethane  foams, the price responsiveness of CFC
use in thermoformed extruded  PS  sheet depends on the  costs of
alternative blowing agents  and CFC recovery, and varies  with plant size.
Unlike flexible  foams, however,  consumer response can be expected to  be
substantial here.  That is  because packaging materials using wood fiber
compete closely with similar  products made  from  PS  sheet.
     Our analysis distinguishes  among small, medium,  and large plants,
consuming 350,000, 500,000, and  750,000  pounds of CFC-12 per year,
respectively.  Extruded PS sheet accounts  for  about 4.7  percent of total
use of CFC-12 in the United States.2  Table 4.1  shows the prices,  in
     2For more information on the cost assumptions underlying our
analysis, see Appendix B.

-------
                                  - 22  -
 1985 dollars,  at which these plants should voluntarily reduce
 their use of CFG-12.  The  format  for this table  is the same as
 that in Table  3.1.
     Two patterns are immediately apparent.  First, larger plants always
 act to reduce  their use of CFC-12 at lower prices than smaller plants.
 Second, it appears that any size  plant would prefer to convert to
 pentane than to recover and recycle CFC-12.  This second point suggests
 that recovery  and recycle will occur only if conversion to pentane is
 not possible;  recovery and recycle options are shown below the dotted
 line to emphasize their conditional nature.  Concern over pentane's
 contribution to smog, noted above, could make conversion difficult.  The
 reduction numbers shown here for  recovery and recycling assume that, for
 whatever reason, no conversion occurs.  These reduction numbers do not
 reflect substitution toward wood-fiber-based packaging materials if the
 price of CFC-12 rises.  Such substitution could be substantial.
 Although good  empirical data on the elasticity of demand for CFC-based

                               Table 4.1
          TECHNICAL OPTIONS TO REDUCE THE USE OF CFC-12 IN THE
            MANUFACTURE OF THERMOFORMED POLYSTYRENE SHEET[a]
Price
0.62
0.79
0.99
0.80
1.20
1.70
Plant
Size
Large
Medium
Small
Large
Medium
Small
Action
Convert
Convert
Convert
Recycle
Recycle
Recycle
% of
Market
25
50
25
25
50
25
% Reduction
in Plant Use
100
100
100
50
50
50
Total %
Reduction
25.0
50.0
25.0
12.5
25.0
12.5
   [a] See the text for an explantion of the options shown below the
dotted line.

-------
                                 - 23 -
packaging material do not exist, most industry observers agree that wood
fiber and plastic are close enough substitutes to make this elasticity
very high.  The actual value of this elasticity is not relevant to
policy unless foamers cannot convert to pentane, because conversion is
cost-effective at such a low price that foamers stop using CFC-12
quickly, leaving little opportunity for consumer response to affect its
use.
     The current price of CFC-12 is $0.67 per pound, suggesting that
large plants should already have converted to pentane.  In fact, as
noted above, pentane is a competitive substitute for CFC-12.  We do not
know exactly what share of the market currently uses pentane; we can
resolve this by considering two alternative base cases, one without any
conversion and another with total conversion among large plants.  This
is how we treat the reduction options shown here in Sec. VIII.
     Technical control options, then, can easily eliminate the use of
CFC-12 in the manufacture of thermoformed PS sheet.  But this is a small
use of CFC-12.  Even total elimination of this market would cut the use
of CFC-12 in the United States by only 4 to 5 percent.

-------
                                 - 24 -

                            V.   SOLVENTS
      In  the United States, CFC-113 and methyl chloroform are used
primarily as solvents.  About 85 percent of CFC-113 production is used
to  clean and dry electronic, metal, and plastic parts.  About 70 percent
of  methyl chloroform is used to degrease and clean metal and electronic
parts.   We focus here on technical control options that could reduce the
use of these PODs in cleaning and drying applications; they fall into
three categories:  vapor recovery, recovery from waste, and product
substitution.  We have quantified the social cost and effectiveness of
these options for reducing the use of CFC-113 and the discussion below
focuses  on this chemical.  We have not examined methyl chloroform in as
great detail, but believe that similar measures could be applied to this
POD with similar effectiveness and somewhat higher social costs.1

VAPOR  RECOVERY
     Vapor recovery can be effected by means of improved equipment
designs  or carbon adsorption techniques.

Improved Equipment  Design
     Improved equipment designs include increasing the freeboard height
of  the equipment (the height of the equipment wall above the solvent
surface), using refrigerated condensing coils, adding a freeboard
chiller, and reducing the throughput speed of items being cleaned to
allow more time for the solvent to drain back into the equipment.
     1Carbon tetrachloride may be an important solvent in the developing
world but plays no such role in the United States.  In the near future,
almost all U. S. carbon tetrachloride production will be used to produce
CFC-11 and -12.  As a result, we do not discuss it here.  However,
technical control options like those discussed here may be relevant to
some uses of carbon tetrachloride outside the United States.

-------
                                 -  25  -
Carbon Adsorption Techniques
     Carbon adsorption techniques employ specially prepared carbon beds
over which solvent-laden air is passed.   The solvent  adsorbs onto the
carbon, after which it is desorbed,  usually with  steam,  and recycled.
If mixtures of solvents are used rather  than pure CFC-113, the soluble
components of the mixture will at least  partially be  retained in the
water in the steam desorption step.   In  this case, the original
proportions of the solvent mixture must  be reconstituted before it can
be reused and the water phase presents disposal problems of its own with
associated costs.

RECOVERY FROM WASTE
     Purification of CFC-113 from the waste liquids that are removed
from vapor degreasers or cold cleaning units is accomplished through one
or both of the following methods:  in-house distillation techniques, and
external-to-the-plant reclamation services.

In-House  Distillation
     Distillation units boil off, condense, and  collect  purified  solvent
from the waste liquids that have been removed from a  piece of  equipment,
leaving behind contaminants such as oil, grease,  and  bits  of  solid
material and flux from the cleaned components.  If nonazeotrope mixtures
of two or more solvents are used, each component  of £he  mixture will
boil off at a different temperature, necessitating the additional step
of reconstitution of the original proportions of the  mixture.
     One in-house distillation unit is able simultaneously to supply
purified,  reclaimed solvent to several vapor degreasing machines.
Because of this, in-house distillation is most attractive  for users  with
multiple degreasers.

External Reclamation
     About 20 percent of cleaning and drying solvent  losses2  end  up  as
     'Cleaning and drying applications represent about 84 percent of
CFC-113 total use.

-------
                                 - 26 -
wastes that could be sent to external reclamation services.  Not all of
this waste would be reclaimed, since some is too contaminated to warrant
reclamation and some cannot be extracted from the waste even with good
distillation.  Moreover, some waste from small users accumulates slowly,
making outside reclamation uneconomical because of the costs of
collection and transportation.  Finally, because reclamation is largely
an option only in cleaning and drying applications, even if as much as
90 percent of the waste could be returned to use, total CFC-113
emissions could be reduced by no more than 15 percent.

PRODUCT SUBSTITUTES
     The most commonly mentioned substitutes for CFC-113 are methyl
chloroform (which is itself a potential ozone depleter), methylene
chloride, trichloroethylene, perchloroethylene, and in some cases,
deionized water.  Except for water, however, none of these substitutes
is as gentle a solvent as CFC-113.  They may affect some plastics or
some of the delicate materials in electronics components and are not
suitable for certain specialty dry cleaning operations, such as the
cleaning of leather.  In addition, none of the substitutes except water
has as high a threshold limit value as CFC-113, and thus more care would
have to be taken with the substitute solvents to see that their levels
in the workplace remain within Occupational Safety and Health
Administration health standards.3
     A ban on land disposal of chlorinated solvents was mandated in the
1984 amendments to the Resource Conservation and Recovery Act.  These
amendments, which will go into effect in November 1986, might induce
some solvent consumers to adopt alternatives, rather than pay for more
expensive disposal options such as incineration.  Incineration of
CFC-113 is likely to be more costly than incineration of the other
chlorinated solvents.   CFC-113 contains fluorine, which industry sources
indicate deteriorates  some of the refractory materials lining
incinerators.   This might constitute an additional incentive for
substitution of other solvents for CFC-113.
     3In general, the higher the threshold limit value, the less
hazardous the substance.

-------
                                - 27 -
QUANTITATIVE ANALYSIS OF TECHNICAL CONTROL OPTIONS
     Our quantitative  analysis of technical options to reduce the  use--
and hence emissions--of  solvents starts by recognizing that CFC-113  is
not embodied in the  final  products that create the demand for CFC-113.
To reduce the use of this  chemical, we must reduce the CFC-113 vapor and
waste losses that occur  during the cleaning or drying process.  We
identify six market  segments  in which these losses can be reduced, based
on three sizes of units  and on whether the CFC-113 is used in pure form
or mixed with other  chemicals.  Hence, our analysis focuses on 12  places
where actions might  be taken  to reduce losses.  Table 5.1 indicates
their relative importance  in  the U.S. market for CFC-113.  The table
indicates that losses  are  most important  in small units and that vapor
losses are more important  than waste  losses.  This is true in part
because the owners of  larger  units have already taken actions to reduce
losses and waste losses  are easier to reduce than losses from vapor.
                               Table 5.1
      WHERE LOSSES CURRENTLY OCCUR IN THE  U.S. MARKET FOR CFC-113
                        (Percent of total  loss)
Type of
Loss
Vapor losses with
pure CFC-113
Waste losses with
pure CFC-113
Vapor losses with
combined solvent
Waste losses with
Small
(15 Gallon)
Unit
16.0
8.0
16.0
8.0
Medium
(60 Gallon)
Unit
11.7
3.3
11.7
3.3
Large
(375 Gallon)
Unit
9.6
1.4
9.6
1.4
  combined solvent

-------
                                  - 28  -
     Within the appropriate sectors, our analysis considers the
 opportunities  for reducing losses by using more conservative equipment,
 recovering and recycling vapor  losses; distilling waste CFC-113
 in-house, reclaiming waste CFC-113 externally, or substituting toward
 alternative solvents.  We model the first three options by comparing the
 costs of providing a given level of solvent services in each sector
 under different arrangements.  We assume that users increase their use
 of external reclamation and switch more toward other solvents as the
 price of CFC-113 rises.4  Table 5.2 summarizes the results of this
 analysis.5  Because it is a complicated table, we review its structure
 briefly before considering the  implications of the numbers reported in
 it.
     The first column lists prices of CFC-113 at which changes occur.
 The second column indicates where the changes occur.  The third shows
 the action taken at each price.  "Recover" refers to recovery and
 recycle of vapor losses, "conserve" to a purchase of more conservative
 equipment, "new solvent" to a decision of some fraction of CFC-113 users
 to replace CFC-113 with another solvent, "reclaim" to external
 purification,  and "distill" to the use of in-house distillation
 equipment to recover waste.6  The fourth column shows the share of the
 total market for CFC-113 affected by these changes.  The last column
 reports the total incremental percentage effect of each change on the
 demand for CFC-113 in drying and cleaning.  The fifth column shows the
     *For more detail on the analysis of technical options for reducing
the use of CFC-113, see Appendix B.
     5Table 5.2 summarizes results for the "High Quantitative" case in
Sec. VIII, which assumes that none of the options that are cost-
effective at the current price of $0.89 per pound have been exploited.  A
similar table underlies the "Low Quantitative" case, but is not shown
here.
     6In each case, with one exception, the action refers to a decision
to add a measure to those already undertaken at lower prices.  The one
exception occurs in medium size units using pure solvent.  Here, the
user will adopt more conservative equipment only in a price range from
$0.89 to $1.16.  The user adopts recovery and recyle only for prices
between $1.16 and $4.46.  For prices above $4.46, the user adopts both
more conservative equipment and recovery and recycle technology.

-------
                              - 29 -
                            Table 5.2

  TECHNICAL OPTIONS TO REDUCE THE USE OF CFC-113 IN CLEANING AND
                       DRYING APPLICATIONS
Price
0.89
1.00

1.13
1.16
1.27
2.00

2.11
2.56
3.00

3.46
4.00

4.46
5.00

Unit
Type
Large pure
Medium pure
Medium comb.
All
waste
Small pure
Medium pure
Large comb in.
All
waste
Medium comb in.
Small comb in.
All
waste
Small comb in.
All
waste
Medium pure
All
waste
Action
Taken
Recover
Conserve
Conserve
New solvent
Reclaim
Recover
Recover
Recover
New solvent
Reclaim
Distill
Conserve
New solvent
Reclaim
Distill
New solvent
Reclaim
Conserve
and recover
New solvent
Reclaim
% of
Market
9.6
11.7
11.7
100.0
25.4
16.0
11.7
9.6
100.0
25.4
3.3
16.0
100.0
25.4
8.0
100.0
25.4
11.7
100.0
25.4
?o Reduction
from Action
80
51
51
1.5
2.0
80
28
80
25
19
32
23
6.7
4.7
18
2.9
1.2
3
1.6
.8
Total %
Reduction
7.7
6.0
6.0
1.5
.5
12.8
3.3
7.7
24.6
4.8
1.0
3.6
6.7
1.2
1.4
2.9
0.3
0.3
1.6
0.2
NOTE:  For an explanation of the terms used in the table, see the text.

-------
                                 - 30 -
implied percentage effect on the relevant sector.7  Each row identifies
a specific action.  Because the effects of switching to other solvents
and reclaiming externally rise as price rises, their incremental effects
over certain price ranges are aggregated and reported at specific
prices.  For example, switching and reclamation reported at $4.00
includes all of the effects of these actions over the range from $3.00
to $4.00.  In each case, switching accounts for two-thirds of this
aggregation for waste losses and reclamation for the remaining third;
only switching affects vapor losses.  As in earlier tables, we stop
arbitrarily at $5.00.
     Notice first that large reductions occur at lower prices and that
incremental reductions get very small as price rises.  That is because
the level of the demand of CFC-113 falls as price rises, making it
difficult for actions that occur at higher prices to have much effect
relative to the current size of the market.  At $5.00, the market has
fallen to 6 percent of the baseline, suggesting that actions induced at
prices higher than this have little policy relevance.
     Notice next that in-house distillation is induced in only two
market segments.  That is because our analysis indicates that in-house
distillation is so cost-effective that users in all other segments would
adopt in-house distillation at prices well below the current price of
$0.89 a pound.  In fact, in-house distillation is widely practiced in
industry, so this is not surprising.  Nonetheless, it might be argued
that some opportunities for in-house distillation have not been
exploited in the segments excluded here, suggesting that we understate
the opportunities for reduction.  Including these opportunities,
however, would not change the aggregate results significantly.  It would
suggest that distillation in these segments could cut total use about 10
percent.  But CFC-113 saved through distillation would no longer be
"available" to be saved through switching and external reclamation;
     7In earlier tables we calculate the last column from the fourth and
fifth columns.  In this case, some options reduce CFC use in all 12
sectors and weighted sums of the changes in all sectors must be
calculated to determine the total effect on CFC use.  The number
reported in the fifth column here is essentially an average percentage
change in all the sectors affected.  If only one sector is affected, of
course, it is the actual change within that sector.

-------
                                 - 31 -
reductions in the effectiveness of these options would offset the
effects of distillation, leaving the total effect on CFC-113 use only
marginally larger.  Whether we include or exclude these highly cost-
effective reduction opportunities, then, makes little difference.
     Options associated with vapor losses have much larger effects than
those associated with waste losses.  In the price range shown, recovery
and recycling with carbon adsorption reduces use over 30 percent and  use
of more conservative equipment over 15 percent.  All effects on waste
losses, on the other hand, sum to about 10 percent.  That is true in
part because vapor losses currently account for most losses and because,
in this analysis, most segments have already exploited the effective
strategy of adopting in-house distillation.  We use a rather arbitrary
assumption to calculate the effects of external reclamation, but
alternative assumptions within a reasonable range would not change the
qualitative results reported here.*
     Decisions to switch to an alternative solvent shrink demand by
about 30 percent over the price range shown, accounting for a
substantial share of the reductions here.  Our assumption about how
users respond to rising prices could significantly affect this result
and hence the aggregate results shown.9  Hence better information on the
empirical adequacy of that assumption would be desirable; it  is not
available at this time.
     In the end, the most important implication of Table 5.2  is that
almost all use of CFC-113 can be eliminated in drying and cleaning
applications at an incremental social cost of  little more than $4.00 per
pound ($5.00 - $0.89).  Although that cost is high relative to the
current price of CFC-113, it is much lower than the cost one  might
     *We assume that reclamation shrinks waste losses by 0.5 percent of
actual losses for every 1 percent rise in the price of CFC-113 above
$0.89.  Because reclamation accounts for only about a 7 percent
reduction, changing the elasticity from 0.5 to some other number would
not affect the aggregate results much.
     'We assume that, above a certain "switching price," users will
reduce their use of CFC-113 by 1 percent of the actual use for every 1
percent rise in its price.  The switching'prices, $1.05 for small units,
$0.97 for medium units, and $1.21 for large units, reflect the increase
in price above current prices required to recover the incremental
investment needed to make a change in solvent feasible.

-------
                                 - 32 -
associate with implementing other CFC control options, for example, in
refrigeration applications.  Much of the reduction, however, comes from
switching to alternative solvents.  Using price incentives to induce
this would be more effective than using mandatory controls because of
the detailed analysis required to determine where switches can occur,
and to what alternative, without significantly degrading the cleaning
and drying function.  Again, cleaning and drying accounts for about 85
percent of the total U.S. use of CFC-113, suggesting that the options
shown here could significantly reduce total use of this chemical.   Other
options are available to reduce CFC-113 use in other applications  but,
because of the small size of these individual applications, they are not
discussed here.
     Although we have not studied it in detail, we believe similar
techniques could be used to reduce U.S. use of methyl chloroform in
solvent cleaning operations.  Options associated with external
reclamation and switching to alternative solvents could occur at costs
similar to those presented here;10 the costs of other options, including
in-house distillation and carbon adsorption, would be higher than  those
shown here.  Like the use of CFC-113, then, the use of methyl chloroform
in drying and cleaning could be effectively eliminated in this country
at a reasonable incremental social cost.  This includes about 70 percent
of the U.S. market for methyl chloroform.  Options not discussed here
could be used to reduce the use of methyl chloroform in certain other
individual applications.
     10CFC-113 and methyl chloroform are potential substitutes for one
another in some applications.   We do not address such interactions here.

-------
                                 - 33  -
            VI.  REFRIGERATION AND AIR CONDITIONING
     Refrigeration and air conditioning  applications employ CFC-11 and
CFC-12, as well as other chemicals.   CFC-11  is used in commercial and
industrial air conditioning systems  (chillers), and CFC-12 is employed
in mobile air conditioning units,  chillers,  retail food refrigeration
systems, and home refrigerators  and  freezers.
     Many options have been identified for reducing CFC emissions from
refrigeration and air conditioning applications.  Those for which we
have sufficient information to formulate price-reduction pairs are:
substitution of chemicals used to  locate system leaks at manufacturing
and installation and used as refrigerants; and recovery at disposal.
     After discussing these options,  we  turn to other potentially
important options for which our information  is not so complete.

OPTIONS DOCUMENTED WITH QUANTITATIVE ANALYSIS
Chemical Substitution:   Test Gases for Locating  System Leaks
     Manufacturers and installers  of chillers and retail  food
refrigeration units can substitute the refrigerant gas CFC-22 for CFC-12
in leak testing applications.  CFC-22 is a POD, but  its potential effect
per gram on stratospheric ozone is so much smaller than that  of  CFC-12
that substitution would be worthwhile.   Halide "sniffers," which are
used widely throughout the industry, could effectively detect leaks of
the halogen.  Substituting CFC-22  (or another gas) can be inconvenient
for an installer, however, who charges the  refrigeration  unit on site
(as is done with retail food refrigerators).  The convenience to the
service person of carrying one or  more cylinders  containing  only one
refrigerant, using that both to test and to  charge the system,  and
avoiding one additional evacuation of the test gas is  influential  in  the
service person's choice of test gas.  The possibility of  contaminating
the CFC-12 with the CFC-22 test gas  is also  a drawback.   The two CFCs
form an azeotrope that is difficult to separate.

-------
                                 - 34 -
Chemical Substitution:   Refrigerant Replacement of CFC-12 with  CFC-502
     CFC-502 is an azeotrope composed of CFC-22  and CFC-115.   It is
considered to pose a far smaller potential threat  to the  stratospheric
ozone layer than CFC-12.  Low temperature retail food applications today
almost exclusively employ CFC-502.   Its use in medium temperature units
has been increasing in recent years, possibly because of  the convenience
of using one refrigerant for both applications.  Furthermore,  CFC-502
requires a smaller, less costly compressor.   The disadvantage  of CFC-502
is that it is almost three and a half times more expensive than CFC-12.
If the price of CFC-12 were to rise enough,  there  would be a strong
incentive to use CFC-502 in all new medium temperature freezer
applications.

Recovery at Disposal
     Recovery is already practiced to some extent  in chiller and retail
food applications.  The design of most home refrigerators and  freezers
does not allow recovery and the small amount of  CFC-12 that could be
recovered from them is unlikely to warrant redesign.   The major
unexploited opportunity to recover CFC-12 at disposal is  in mobile air
conditioning units.  Motor vehicle salvage yards recover  a wide variety
of materials, and they might also collect used CFC if there were a large
enough market for it.

OTHER OPTIONS  TO REDUCE POD EMISSIONS
     For some potentially important technical control options, our
information is more qualitative than for the options discussed above.
These options are ranked in the order of their estimated  potential
impact on POD use.

Chemical Substitution Using New Gases under Development
     FC-134a is a fluorocarbon that contains no  chlorine, and  thus poses
no known threat to the stratospheric ozone layer.   The chemical  can be
substituted for CFC-12 in existing refrigeration and air  conditioning
equipment with little or no redesign, although it  does require that a
different oil be used with it.  Refrigeration and  air conditioning

-------
                                 - 35  -
applications in 1984 accounted for 37 percent to 74 percent  of  total
U.S. CFC-12 use.1  To date,  however,  no commercial manufacturing process
for FC-134a is known to be available.  Until such a process  is
developed, it does not offer a viable option for reducing the use of
CFC-12.

Reduction of Emissions during  Servicing
     Servicing emissions could be reduced if refrigeration and  air
conditioning units were pumped down before servicing,  and the recovered
refrigerant decontaminated if necessary and reused.  About 16 percent of
the total CFC-12 used in the United States in 1984 was lost because of
venting during servicing of mobile air conditioners.  CFC-11 and CFC-12
losses in 1984 during servicing of chillers amounted to 1.3 percent and
0.6 percent of the total U.S. use of those chemicals during that year.
Losses from retail food refrigerators during servicing amounted to 0.8
to 8.7 percent of U.S. CFC-12 use during 1984.2

Equipment Replacement
     Reciprocating and screw compressor chillers using the refrigerant
CFC-22 have the potential to compete with centrifugal chillers in the
200, 300, and possibly 400 ton range.  By replacing centrifugal machines
in these ranges, CFC-11 and CFC-12 use in all centrifugal machines could
be reduced by about 25 percent, which in 1984 would have corresponded to
about a 1 percent reduction in combined U.S. CFC-11 and CFC-12 use.
      *In work elsewhere, Rand identifies 37 percent of CFC-12 as being
used  in these applications in the United States.  The applications of
another 42 percent of CFC-12 use cannot be identified.  The available
data  suggest that as much as 88 percent of this unexplained portion may
be used for refrigeration and air conditioning.  For details, see
Hammitt et al.  (1986).
      2As indicated in the footnote above, how much CFC-12 is used for
refrigeration applications is uncertain.  Retail food refrigerators
account for 3 to 33 percent of all U.S. CFC-12 use.

-------
                                 - 36 -
QUANTITATIVE ANALYSIS OF TECHNICAL CONTROL OPTIONS
     Our analysis of control options relevant to air conditioning and
refrigeration is simpler than the analyses underlying other results
reported here.  For two options, we compare the prices of CFC-22 and
-502 with that of CFC-12; when relative prices have changed
sufficiently, users switch from CFC-12 to the relevant alternative.  For
the third, we use a simple cash flow analysis to determine at  what price
it would pay the owner of a typical salvage yard to invest in  recovery
equipment.  We draw heavily on the methods and data reported in Palmer
et al. (1980, pp. 141, 158, 185-189) to make these calculations; for
methodological details, see that reference.
     Table 6.1 summarizes the results of this analysis.   Its format  is
similar to that used above.  Because we do not look at individual market
segments within each application, the fourth column shows the  share  of
each application in total U.S. use of CFC-12.  The final column shows
the effect on total U.S. CFC-12 use.  A brief explanation of the result
in each row follows.

                               Table 6.1
          TECHNICAL OPTIONS TO REDUCE THE USE OF CFC-12 IN AIR
              CONDITIONING AND REFRIGERATION APPLICATIONS
Price
1.00
1.17
3.64
Application
Retail food
Retail food
Cent, chiller
Mobile air cond.
Action
Taken
Use CFC-502
Use CFC-22
Use CFC-22
Recover
% Share
of Use[a]
3-33
3-33
1.1
32.6
% Reduction
in
Application
48.0
5.0
1.4
2.5
Total %
Reduction
1.4-15.6
0.15-1.6
0.02
0.82
   NOTE:  See the text for an explanation of the abbreviated expressions
used in the table.
   [a] This range results from uncertainty about how CFC-12 is used in the
United States.  For details, see Hammitt et al.  (1986).

-------
                                 - 37 -
Chemical Substitution  of CFC-12 with CFC-502
     The first row shows  that a CFC-12  price increase  to  $1.00 per
pound, which reduces the  price differential between CFC-12  and CFC-502
by 20 percent, would lead to all new medium temperature retail store
refrigeration systems  being charged with CFC-502.   Because  of the
convenience of using a single refrigerant for leak testing  as well  as
charging, implementation  of the above option would probably eliminate
some or all use of CFC-12 during installation of retail systems.  A
conservative estimate is  that half of all CFC-12 leak  testing would be
replaced by the use of CFC-502.  Although switching to CFC-502 as the
charge installed in new medium temperature retail food units has a  large
effect on this application of CFC-12, this application may  account  for
only a small share of total CFC-12 use.  Rand has identified only 3
percent of CFC-12 as being used in retail food refrigeration.  Evidence
suggests, however, that that estimate could be ten times  higher,  in
which case this substitution would have a substantial  effect.

Chemical Substitution  of  Test Gases
     CFC-22 can be substituted for CFC-12 on a pound-for-pound basis  for
leak testing at manufacture in centrifugal chillers and for leak testing
at manufacture and installation in retail refrigeration units.   Because
CFC-22 currently costs $1.17 per pound in real terms,  the switch to
CFC-22 as the test gas would occur if the CFC-12 prirce rose $0.50 above
its current bulk price of $0.67.  The first row at a price of $1.17
shows the remaining half of leak testing in retail food refrigeration
units switching from CFC-12 to CFC-22.   The second row shows a similar
shift in centrifugal chillers.  In both cases, the effect on the
particular application of CFC-12 is small and the shares of these
applications in total U.S. use of CFC-12 are also small.

Recovery at Disposal
     The last row indicates that motor vehicle salvage yard operators
would find it cost-effective to recover CFC-12 from mobile air
conditioning units at a CFC-12 price of about $3.64 per pound.  Despite
the importance of mobile air conditionaing to total CFC-12 use, however,
this option has a small aggregate effect.

-------
                                 - 38 -
     In sum, few options are available to reduce the use of CFC-12 in
air conditioning and refrigeration at a reasonable social cost.   That is
true either because the applications where options exist account for a
small share of total use of CFC-12 or because the options that are
available have only small effects where they are applied.  It is worth
noting that a large portion of current CFC-12 cannot easily be
associated with any particular use.3  But even if all of the unexplained
portion of CFC-12 use in the United States were in fact associated with
air conditioning or refrigeration applications, this would not change
the basic qualitative conclusion that flows from this table.  Only if a
new chemical, like FC-134a, can replace CFC-12 in a wide variety of
these applications can we expect to see a significant reduction in its
use here at reasonable social costs.
     3For a discussion of this point, see Hammitt et al. (1986).

-------
                                 - 39 -
                VII.   MISCELLANEOUS APPLICATIONS
     Potential ozone depleters  are used in a number of smaller
applications in addition to  the major categories discussed above.
Technical options available  in  some  of these smaller applications are
probably effective enough to induce  significant reductions in the total
use of CFC-11 and -12 and Halon 1301.  This section addresses these
options.  We have not quantified  the effectiveness of using these
options to reduce the use of PODs nor have we estimated their social
costs; these are subjects for future research.  But these options could
be as cost-effective as the  options  considered above.
     We consider options that can affect the use of CFC-11 and 12 in
exempt aerosol uses, liquid  food  freezing, sterilants, and Halon 1301 in
fire extinguishing systems.

EXEMPT AEROSOL PROPELLANT  USES
     Most uses of CFC-11 and -12  as  aerosol propellants were banned in
the United States in 1978.  In 1985, about 9500 metric tons  (mt) of
CFC-11 and CFC-12, or 3.9 percent of their total use in the United
States, were used in exempt  applications, including certain
insecticides, military uses, and  some medicinal uses.  Generally, these
exceptions were granted because the  alternative propellants that were
available at the time the ban was promulgated were unsatisfactory.
     Since that time, propellent  technology has advanced  considerably.
Although this in itself has  not caused  a  substantial move away from
exempt applications of CFC-11 and CFC-12, some substitution has  occurred
and more could be expected if the prices, of CFC-11 and -12 rose  or the
definitions of exempt uses were tightened.  Two technological changes
are particularly important.

-------
                                 - 40 -
Dimethyl Ether/CFC-22  Blends
     One recent technological advance has been the introduction of
dimethyl ether/CFC-22 blends under the DuPont trade name of Dymel®.1
This blend is a nonflammable alternative to CFC-11 and -12 that can be
used where fire hazards have precluded substituting hydrocarbon
propellants for CFCs.  For example, naval shipyards often have local
rules prohibiting flammable propellants.  Since the uses are qualified
as military uses, the shipyards can (and do) demand that CFC propellants
be used.  Carbon dioxide is a nonflammable alternative, but it has had
other disadvantages that users have been unwilling to accept.  Dimethyl
ether/CFC-22 provides a better nonflammable alternative, and at least
one shipyard has accepted this type of propellant.
     Because Dymel® is more expensive than CFC-11 and CFC-12 in most
applications, most manufacturers have not switched.  But higher prices
for CFC-11 and -12 could promote substitution toward Dymel® in certain
exempt uses.  The current price for a CFC-11/12 blend is about $0.60  a
pound, whereas an equivalent amount of Dymel® currently costs about
$0.90.  Hence, a 50 percent increase in CFC prices would make the two
options far more competitive.  For lack of good data on the specific
uses of CFC-11 and -12 as aerosol propellants, we cannot estimate how
many of these exempt uses Dymel® might be able to displace.  A rough
guess might be about half.

Improved Technologies for Carbon  Dioxide
     The technology for using carbon dioxide as an aerosol propellant
has also advanced considerably.  The original objections to carbon
dioxide centered on poor atomization, varying pressure over the life  of
the can, changing spray delivery rates as the can was used, and the
occasional inability to expel the entire contents of the can, as a
result of loss of pressure.  New mechanical break-up valves have solved
many of the atomization problems, and advances in formulation and
filling technology have reduced the other problems.  As a result, carbon
     1 Dymel® trademark covers a variety of dimethyl ether/CFC blends.
Of these, a blend with CFC-22 is nonflammable.  Blends with CFC-142b or
CFC-152a are flammable.

-------
                                 - 41 -
dioxide now appears to be a cost-effective alternative for  many aerosol
applications in which it was unacceptable five years  ago.   Many
manufacturers may not have investigated these new options because  they
are not yet aware of the improvements that have been  made in carbon
dioxide propellant technology.   We do not know how important these
applications are to exempt uses of CFC-11 and -12 and hence cannot
estimate how much CFC could be displaced by the use of carbon dioxide in
these app1ications.

LIQUID FOOD FREEZING
     In 1985, about 3000 mt, or 1.9 percent of CFC-12 total production,
were used and emitted in liquid food freezing (LFF) applications.  One
promising option for reducing CFC-12 emissions from LFF is  substitution
of alternative freezing methods.  Industry sources suggest  that one  such
method, the air blast system, is making strong inroads into the LFF
market today.  Many current LFF users will eventually replace LFF
capacity with air blast systems except in certain seafood and fruit
applications (Mooz et al., 1982).  This trend is borne out  by industry
sources who suggest that even at present CFC-12 prices, there is
apparently some movement away from LFF systems.  If the price of  CFC-12
were to rise slightly, we would expect a more rapid shift away from  LFF.

STERILANTS
     In 1985, about 8000 mt, or 5.0 percent of total  CFC-12 production,
were used in a mixture with ethylene oxide for sterilizing applications
in hospitals and institutional settings.  One option that could reduce
CFC-12 use and emissions is recovery and recycle of the sterilant in
industrial sterilization operations, which account for about half the
present use.  In our earlier work, two systems with recovery
efficiencies ranging between 70 and 100 percent were in various  stages
of commercialization (Mooz et al., 1982).  At that time, the economics
appeared attractive, especially for large institutional users.  Assuming
that industrial facilities fully adopt these systems  and maintain a
recovery efficiency of 85 percent, the savings in 1985 CFC-12 purchases
could amount to 3400 mt or about 2.1 percent of total CFC-12 production.

-------
                                 - 42 -
HALON 1301
     Halon  1301 total flooding systems have been used on a large scale
for only about a decade.  Emissions from such systems are small at
present but may become significant because of recent and continuing high
growth.  Current U.S. use amounts to about 5400 mt.
     One option that could reduce future emissions is substitution of
alternative extinguishant systems; candidates include water,  foam, and
carbon dioxide.  Although each of these is less costly, the Halon system
offers the advantage of leaving no residue on expensive electronic
equipment in case of a fire.
     A second option is recovery at disposal.  Industry sources claim
that the high cost of Halon 1301--several dollars a pound--ensures that
it is already recovered at disposal or remodel.  One installer, however,
indicates that systems under about 60 pounds are commonly vented.   There
is little experience with disposal of such systems and future research
will investigate such practices further.

-------
                                 - 43 -
     VIII.   CONCLUSIONS AND DIRECTIONS FOR FUTURE WORK
     The results developed above on the costs  and effects  of  specific
technical options provide the building blocks  for policy analysis  that
examines alternative ways to reduce the emissions of potential  ozone
depleters.  We plan to use these results to examine policy alternatives
in the future.  This section summarizes our results and suggests
directions for detailed future analysis.  It uses the results developed
above to compare the relative price responsiveness of CFC-11, -12,  and
-113.  It summarizes the total effects on the  use of these chemicals  if
all of the options reviewed here were successfully implemented.  And  it
compares the results developed here with those developed in Palmer et
al. (1980).  This comparison confirms the need for an update just  five
years after Palmer et al. was published, suggesting that we should be
cautious about applying the results presented  here or those in  Palmer et
al. to forecasts in the future.
     Palmer et al. (1980) developed demand schedules for CFC-11,  -12,
and -113 by looking at a set of technical control options  similar  to
those considered quantitatively in this analysis.1  It developed
schedules for 1980 and 1990.  The analysis performed here looks at the
costs and likely effects of these options again, using 1985 as  a
reference year.  But our analysis deals with these options in somewhat
different ways and'considers others where the quantitative data are not
so complete.  These differences in approach, coupled with our access  to
more up-to-date information, explain the differences between the  results
in Palmer et al. and those developed in this update.
     There are three significant differences between the analysis
performed here and that in Palmer et al.  First, the analysis here
identifies control options that appear to be cost-effective at  current
     1The demand schedules for Palmer et al. are summarized in a table
on p. 265 in that document.  The schedules shown here for Palmer et al.
are based on the data "in that table and the price levels assumed in the
analysis:  $0.34 per pound for CFC-11, $0.41 per pound for CFC-12, and
$0.62 and $0.40 per pound for CFC-113 in 1980 and 1990, respectively.

-------
                                  - 44  -
 prices.  Discussions with industry confirm that portions of the industry
 have  adopted  a number of these options.  But we do not know how much
 potential  remains  for these options; hence, we develop a range of
 results to bracket the potential that still exists.  This issue does not
 appear in  Palmer et al.
      Second,  although the results in Palmer et al. reflect the fact that
 the uses of significant portions of total CFC-11 and -12 production have
 not been satisfactorily explained, that report does not attempt to
 suggest where those uses might be.  We consider alternative ways in
 which the  unexplained portion of the production of these chemicals might
 be used and their  implications for the effects of control options.
     Finally, Palmer et al. focused solely on control options for which
 quantitative  data were available on effects and costs.  This makes the
 estimates  of  effects in Palmer et al. conservative.  We consider other
 options and use some simple rules to suggest what their effects might
 be.
     The net  result of these changes is that the results reported here
 provide a  wider range of outcomes than those reported in Palmer et al.
 Although Palmer et al. can marshal more empirical data to support the
 specific results reported, our ranges are more likely to include the
 actual effects of the control options that we believe are available to
 reduce the use of CFC-11, -12, and -113.  Consider the results for each
 chemical in turn.
     Figure 8.1 presents demand schedules for CFC-11.  The vertical axis
 shows percentage increases in price; the horizontal axis shows
 percentage reductions in use associated with these price increases.
 Hence, the schedules are "inverted" demand schedules of the type
 discussed  in Sec. II.  The figure shows five schedules.  Those shown as
 solid lines are based on material in this Note; the dotted schedules are
 based on results from Palmer et al.
     The three schedules based on the analysis in this Note reflect
different  assumptions about what options are available to reduce the use
of CFC-11.   The most conservative schedule is the "Low Quantitative"
 schedule.  It includes only the effects relevant to the one use for
which we have quantitative information--flexible slabstock foam blowing--
 and assumes that all options that appear cost-effective at current

-------
                                 - 45 -
                   5  1    10       15       20       25
                       Percentage reduction in CFC-11 use
          Fig. 8.1 -- Demand schedules for U.S. use of CFC-11
prices have already been exploited.  The "High Quantitative" schedule
also focuses on slabstock foam blowing, but assumes that no large plants
have moved to the 75 percent use of methylene chloride suggested by the
first option in Table 3.1.  Hence, this option is still available to
reduce the use of CFC-11.  The "Total" schedule adds options relevant to
molded foam, exempt aerosol uses, and uses in chillers.  We do not know
the exact prices at which users would adopt these options; we adopt the
simple rule of allocating a quarter of the total potential reduction

-------
                                 - 46 -
 from using these options to each price increment from 100 to 200
 percent, 200 to 300 percent, and so on.  In sum, options relevant to
 slabstock foam blowing offer reduction opportunities that fall somewhere
 between the Low and High Quantitative schedules.  The Total schedule
 indicates the largest reduction we expect to be possible at each price
 level.  Hence, the actual demand schedule should fall somewhere between
 the Low Quantitative and Total schedules shown here.
     The two schedules based on the results of Palmer et al., "Palmer
 1980" and "Palmer 1990," reflect demand schedules for 1980 and 1990.
 Presumably the schedule for 1985, which would be most appropriate for
 comparison with our results, falls somewhere between these schedules.
 The Palmer schedules include options relevant to slabstock and molded
 foam.  These schedules indicate that much higher savings are possible
 here than our results indicate.  That is true for two reasons.  First,
 foam blowers have increased their use of methylene chloride since the
 Palmer et al. analysis was completed, thereby reducing the opportunities
 available to reduce CFC-11 use in the future.  Second, partly as a
 result, flexible foam blowing does not constitute as large a share of
 the use of CFC-11 today as expected in Palmer et al.  Flexible foam
 blowing currently accounts for about 17 percent of CFC-11 use.  Hence,
 the Palmer et al. schedules should not include reductions in the use of
 CFC-11 that exceed this level.  Limiting the flexible foam market to 17
 percent of CFC-11 use would bring the Palmer et al. schedules more in
 line with our Total schedule,  which includes molded foam.
     Our schedules show a significant range of possible outcomes.  But
 all of them point to extremely inelastic demand.  Arc elasticities based
 on a 500 percent price increase range from 0.01 to 0.03.
     Figure 8.2 shows a set of schedules for CFC-12 similar to those
 shown for CFC-11 in Fig. 8.1.   The solid schedules reflect results
 developed here; the dotted schedules reflect results from Palmer et al.
     The three solid schedules bear labels similar to those for CFC-11,
but they differ in important ways.  The Low Quantitative schedule once
 again focuses on quantitative options, this time the extruded
polystyrene sheet foam blowing, mobile air conditioning, chillers, and
 retail food refrigeration options shown in Table 6.1.  It assumes that
 large foam blowing plants have all exploited the option of converting to
 pentane, thereby eliminating the first option shown in Table 4.1.

-------
                                 -  47  -
                        10       15      20      25
                          Percent reduction in CFC-12 use
30
35
          Fig. 8.2 -- Demand schedules for U.S. use of  CFC-12
     The High Quantitative schedule assumes that this  option  has  not  yet
been exploited.  It also examines the possibility  that most of the
unexplained use of CFC-12 production in the United States  is  associated
with retail food refrigeration.  If this  is correct,2  the  effect  of
converting retail food refrigeration units to CFC-502  is ten  times
larger than our quantitative analysis based on  known uses  of  CFC-12
     2For a discussion of this  issue,  see  Appendix A  of Hammitt et al.
(1986).

-------
                                 - 48 -
would indicate.  The High Quantitative schedule also incorporates this
possibility.
     The Total schedule adds reductions for which we have not collected
quantitative information.  These include reductions associated with
exempt aerosol propellant uses, liquid food freezing, sterilants,
chillers, and the servicing of refrigeration applications.  As in Table
8.1, a quarter of the total reduction available from these options is
attributed to each range of price increase:  100 to 200, 200 to 300, and
so on.  Hence, the Total schedule once again represents an upper limit
on the size of reductions we would expect in CFC-12 uses if its price
rose.
     The "Palmer 1980" and "Palmer 1990" schedules once again reflect
demand schedules from Palmer et al. for 1980 and 1990; for practical
purposes, they reflect the same schedule.  These schedules reflect the
same options included in the Low Qualitative schedule.  These schedules
closely resemble the Low Qualitative schedule over much of its price
range.  Because they all reflect the same options and do not consider
the unexplained portion of CFC-12 production, this resemblance is not
surprising.  Two factors account for differences between them.
     First, the Palmer et al. schedules assume that extruded polystyrene
sheet foam blowing accounts for about 9 percent of CFC-12 use when in
fact it currently accounts for only 4.7 percent of CFC-12 use.  This
allows Palmer et al. to attribute about twice the percentage savings in
CFC-12 to options that reduce its use in foam blowing that we estimate.
Because foam blowing has not grown as fast as Palmer et al. expected
relative to other uses of CFC-12, our assumption is currently more
appropriate.
     Second, our analysis assumes that the savings that will occur in
foam blowing occur at much lower prices, than those shown in Palmer et
al.  In this Note, savings associated with conversion to pentane all
occur at price increases of less than 50 percent, giving the Low
Quantitative schedule quick reductions at low price rises relative to
the Palmer schedules.  The Palmer schedules show conversion occurring at
price increases of 300 to 350 percent.  Changes in the relative prices
of pentane, CFC-12, and other inputs since Palmer et al. was written
account for this change.

-------
                                 - 49 -
     Where Palmer et al.  concluded that few opportunities  existed to
reduce the use of CFC-12, then,  our analysis suggests  that significant
opportunities may exist.   Unfortunately, our information on those
opportunities is not particularly good.  We do not know the actual size
of the market for retail food refrigerators that account for so much  of
the opportunity shown here.  And we do not know the prices at which
users of CFC-12 would voluntarily adopt the options included in the
Total schedule.  These issues obviously deserve more careful attention.
No matter which of the three solid schedules one accepts,  however, the
elasticity of demand for CFC-12 is low.  Arc elasticities  of the
schedules shown, based on a price increase of 500 percent, range from
0.01 to 0.07.
     Figure 8.3 presents demand schedules for CFC-113.  The solid "Low"
and "High" schedules are similar to those shown in earlier figures.  To
derive the Low schedule, we assume that all options cost-effective at
current prices are fully exploited.  To develop the High schedule, we
assume that the distillation and conservation options that become cost
effective at prices below $0.89 a pound in Table 5.2 are still
available for exploitation.  The actual schedule lies between these.
The two schedules diverge significantly for small price increases but
tend to converge for larger price  increases.  That is because options
that become cost effective at higher prices have  larger effects  if more
CFC-113 is left for them to reduce.
     The dotted "Palmer  1980" and  "Palmer  1990" schedules  reflect demand
schedules for  1980 and 1990 in Palmer  et al.  The  1985  schedule  relevant
for comparison to our results presumably lies between  these  and  hence
cannot differ much from  them.  Almost  all  of  the difference  between  the
Palmer et al.  schedules  and our own can be explained by our  decision to
include recovery and recycle of vapor  losses  with  carbon  adsorption  as
an option.   It accounts  for 27 to  34 percent  of total  saving of  CFC-113,
only slightly  less than  the gap between our schedules  and the Palmer et
al. schedules.
     Our results appear  to be more or  less consistent  with those
reported earlier for the effects  of other  control  options.   Whichever  we
use, the elasticity of demand  for  CFC-113  is  higher than  that for CFC-11

-------
                                  - 50 -
      500
      400
   eo
   g  300
   £

   1  200
      100
                                          Palmer
                                          1990
               10      20      30       40      50
                           Percentage reduction in CFC-113 use
60
70
80
          Fig. 8.3  -- Demand  schedules  for U.S.  use of CFC-113
or -12, at  least over  the whole  range of price increases considered
here.  Arc  elasticities based on a price increase of 500 percent range
from 0.10 to 0.16.  These elasticities are still small when compared to
those for other commodities  traded in the economy.
     Three  simple differences, then,  play the largest role in explaining
differences in the results  for Palmer et al.  and those for this update.
First, the  use of CFC-11 and -12 in applications where control options
were available did not grow  as fast as Palmer et al. had anticipated,
leading to  a need to reduce  the  estimated effects of these options on
CFC-11 and  -12 use.  Second,  uncertainty about the use of the
unexplained portion of CFC-12 production leads to great uncertainty

-------
                                 - 51 -
about the effects of control options for applications  that  might  explain
this use.  By considering the portion of CFC-12 that might  be  used in
retail food refrigeration applications,  this Note significantly
increases the potential importance of control options  relevant to this
option.  Finally, this Note considers many more control  options than
Palmer et al. considered.  Only one of these, carbon adsorption in the
recovery and recycle of CFC-113, is considered quantitatively. But lack
of detailed information need not lead us to conclude that other options
might not be important.  Examination of these other options increases
the opportunities to reduce the uses of CFC-11, -12, and -113  at  social
costs that fall within the range examined here and in  Palmer et al.
Despite all the specific changes in implications for  individual
chemicals, both Palmer et al. and the analysis shown here point to low
demand elasticities for all three of these chemicals.
     Our summary results strongly suggest the potential  leverage  of the
options for which we have not developed quantitative  results.   They
clearly deserve more careful attention in the future.   A better
understanding of where unexplained portions of CFC-12  are used will help
quantify more clearly options relevant to CFC-12.  Reductions beyond
those possible with the options considered here, quantitative or not,
will require the use of much more costly options.  If such reductions
are deemed necessary, however, we will need more information on them.
     The reduction of PODs, of course, is a transnational problem.  The
results developed here apply only to the United States and, as Appendix
C emphasizes, are unlikely to reflect opportunities available.  The cost
of options to reduce the use of PODs outside the United States needs
additional attention.
     The real purpose for developing results like those presented here
is to use them to compare the social costs of using alternative
policies, if need be, to reduce global ozone depletion in the  future.
We currently plan to use the results presented here to look at policies
that spread reductions in U.S. use of these chemicals across time  in
different ways.  As better  information becomes available on the costs of
options elsewhere, we can take a broader, global perspective.

-------
                                 -  53  -
                             Appendix A

         POTENTIAL PROBLEMS  IN  COMPARING ALTERNATIVE
                    POLICIES TO REDUCE  POD  USE

     The demand schedules defined in this Note have an explicit
interpretation.  They represent opportunities to reduce use of potential
ozone depleters that are currently known and that  are implemented to
minimize the social cost of reducing the use of  any particular chemical
by any particular amount.  It is likely  that many  policies would not
induce the kinds of reductions shown in  these schedules.  Other policies
could easily induce reduction options that  we do not yet know about.1
As a result, the results reported in this Note must be usrd with caution
in comparing alternative policies.   This appendix  reviews some basic
issues that must be considered in using  results  like those in the text
to compare alternative policies.
     Suppose, on whatever grounds, that  policymakers decide to promote
technical options of the kind discussed  in  the text to reduce POD use.
How do they identify the technical options  available or assure that
options they choose are in fact implemented in the private economy?
These problems of administering a regulatory policy draw important
differences between policies based on command-and-control and those
based on the price system.  They also have  important.implications for
the analysis of regulatory alternatives. Environmental policymakers, in
the United States and elsewhere, typically  use command-and-control
programs to implement policy.  In such programs, regulators identify the
specific measures that must be undertaken and then monitor performance
to assure that the private economy implements these measures.  This
allows policymakers to target the effects of their decisions and
     JIn the discussion below, "options" always refer to specific
actions in the private sector like implementing recycle and  recovering
of CFC-113, substituting pentane for CFC-12,  or substituting away  from
polystyrene sheet to materials based on wood  fiber.   "Policies"  refer to
specific government programs that set up mandatory controls,  impose
taxes, or create entitlements.

-------
                                  - 54  -
mitigate  the  effects of those decisions on parties who would be hurt
most  if forced to  reduce their use of substances or activities that may
endanger  the  environment.  It may also give policymakers a greater sense
of control  and predictability because they actively participate in the
choice of technical options and focus their attention on achieving
specific  quantity  changes in activities that they believe may threaten
the environment.
      Environmental policymakers make much less use of the price system
to help them  implement policy.  Price-oriented programs typically tax
activities  that may endanger the environment or create entitlements to
rights to engage in potentially dangerous activities.  These
entitlements  can be bought and sold, subject to varying regulatory
restrictions, by companies and individuals who engage in these
activities.   In these programs, policymakers rely on incentives to
encourage actors in the private sector to discover and implement ways to
reduce potentially dangerous activities.  Where taxes are used,
policymakers  may not even attempt to achieve any particular level of
reduction in  activities that may endanger the environment.  A tax can
reflect the perceived incremental social harm associated with an
activity; so  long as the tax forces actors in the private economy to
recognize this potential harm, their decisions will determine the
socially  desirable level of the activity.  In this case, policymakers
need  only choose the "right" tax rate and assure that the tax is in fact
paid.  Creating entitlements requires policymakers to choose and
regulate  the  total level of the dangerous activity and to assure that
only  companies with entitlements engage in it.  But, as in the case of a
tax policy, entitlement programs free policymakers of the responsibility
of telling  individuals and companies precisely what to do and then
assuring  that they do it.
     When only social cost-benefit analysis is used to judge
alternatives, command-and-control programs can never perform better than
price-oriented programs unless the administrative cost of monitoring an
activity where tax payments or entitlements are required is high.  This
is true in part because price-oriented policies by definition
differentiate among consumers and suppliers with different willingness
to pay for products and inputs. 'Consumers the least willing to pay for

-------
                                 - 55 -

 foam products switch to alternatives first when foam producers must pay
 a tax  or buy an entitlement to use PODs in foam blowing.  Similarly,
 laborers with the  least concern about alternative blowing agents consent
 to work with these alternatives at the lowest wages when taxes or
 entitlements drive up the cost of blowing agents based on PODs.  These
 are precisely the kinds of private decisions favored by social cost-
 benefit analysis, and price-oriented programs promote them without
 requiring any special knowledge on the part of regulators.
 Price-oriented programs also outperform command-and-control programs
 because they create incentives in industry to discover and implement a
 broad  range of subtle adjustments and accommodations that command-and-
 control methods simply cannot discover or enforce easily.  To be
 successful, price-oriented regulators must assure that effective
 enforcement creates the right incentives, but they need no special
 knowledge of technology or consumption patterns in the private economy
 beyond that required to enforce taxes or entitlements.
     Despite these apparent advantages of price-oriented systems,
 environmental policymakers continue to rely on command-and-control
methods.  Where taxes and entitlements truly encourage individuals and
 companies to cut back activities that may endanger the environment, they
 clearly contribute to solving the problem at hand.  But where they
 simply force individuals and companies to pay a fee that they cannot
 avoid  in any reasonable way, taxes and entitlements can be seen more as
 fiscal instruments that raise revenue or redistribute wealth than as
valid  instruments of environmental policy.  For example, finding
substitutes for CFC-12 in home refrigerators is so difficult and the
cost of this CFC is such a small portion of the cost of a refrigerator
that tax and entitlement programs are highly unlikely to change the use
of CFC-12 in home refrigerators.  Why burden the owners of home
refrigerators—who will surely end up paying for new taxes or
entitlements--with an environmental program that simply cannot change
their behavior?  Those who defend the use of social cost-benefit
analysis would reply that such fees are just transfers among
individuals, not real costs to the economy.  Policymakers who typically
favor the command-and-control programs that avoid such transfers
apparently care about more than just the implications of social cost-
benefit analysis.

-------
                                  - 56 -
     These observations strongly suggest that social cost-benefit
 analysis by itself is unlikely to provide all of the answers that
 environmental policymakers require before choosing between various kinds
 of strategies.  With that very important caveat in mind, however, we can
 ask how best to use the demand schedules discussed above to compare
 alternative regulatory strategies.
     Two ideas are important to using these demand schedules.  First,
 different policies must be represented by demand schedules that reflect
 different options.2  That is because different government policies make
 different technical options and responses available and only the options
 available under a particular regulatory policy should be considered when
 analyzing that policy.
     Command-and-control regulators can consider only technical options
 that policymakers know currently exist, options that they can hope to
 learn about, and options that by whatever means the policymakers can
 bring into existence in the future.  Further, unless the users of these
 technologies pass their costs forward in higher product prices, the
 demand schedules cannot reflect the mitigating effects of demand
 response; for example, strategies that subsidize POD-reducing options
 can reduce costs to industry but increase social cost by preventing
 consumers from reacting where they can to avoid the real costs of the
 POD-reducing technology.  Finally, they cannot include options whose use
 is too difficult to enforce; the most obvious examples are technical
 options based on improvements in operating practices that regulators
would have to monitor frequently to assure compliance.
     Price-oriented regulators, on the other hand, can include a much
broader set of options in the demand schedules they use.  In fact, this
 can present an analytic problem because price-oriented policies are
designed to generate information in their implementation that
     2In the discussion that follows, it may be easiest to think of the
"inverted" demand schedule above in which reductions in production and
use appear on the horizontal axis and the total level of reduction grows
as price rises.  Including more technical options and responses in such
a schedule allows more reduction at any price, suggesting that adding
options and responses increases the elasticity of the demand schedule in
question.

-------
                                 - 57 -
policymakers and their analysts cannot have when estimating demand
schedules.  That is, analysts must assume something about the private
sector's ability and willingness to innovate to reduce POD use.   If the
analyst assumes that the private sector will innovate at all, he also
implicitly assumes that price-oriented policies can reduce POD use at a
lower social cost than command-and-control policies.  Price-oriented
policies can also accommodate options based on improved operating
practices and other activities that regulators cannot monitor easily.
And they also benefit from the effects of consumer response to price
increases induced by taxes and entitlements.  Only options where the
collection of taxes or enforcement of entitlement use is difficult
should be excluded from these schedules.  As a result, only if these
options account for a large portion of POD saving and can be effectively
monitored in a command-and-control program can such a program hope to
match or surpass the options a price-oriented .program can exploit.
     The second idea to keep in mind when using demand schedules is that
they play different roles in analyzing command-and-control and price-
oriented regulatory strategies.  For command-and-control strategies,
they basically order the options available under this regime to help
regulators determine a sequence of specific options to use to achieve
any reduction in POD emissions.  For price-oriented strategies, they
have a more behavioral role.  They help policymakers estimate how much a
given tax rate might reduce POD use, or what tax or entitlement price
should be associated with any total restriction on POD use.  That is,
although demand schedules simply help command-and-control regulators
order their thoughts, they help price-oriented regulators do things  that
they cannot do in any other way.
     The behavioral role of demand schedules used to analyze price-
oriented policies may prove useful to regulators even if they anticipate
using only command-and-control policies.  Such demand schedules allow
regulators to set a benchmark; regulators can use them to measure the
social cost of reducing a given amount of POD use if the least socially
costly price-oriented program were used.  Policymakers can then compare
the social cost of this alternative with the cost of various command-
and-control programs to determine whether these programs offer enough  to
make them worth the amount they add to the  social cost of  any reduction.

-------
                                  -  58  -
This  is  a comparison that allows policy decisions to be made without
reference to any knowledge about the social benefits of the POD
reduction.
      In  the end, this form of cost-effectiveness is probably the best
form  of  analysis to pursue with the information at hand.  Without good
information on the benefits of restricting the production and use of
PODs, this approach still allows us to compare the social costs of
achieving any given level of reduction in different ways.  The
alternatives to consider can go far beyond the somewhat stark price-
versus-command dichotomy emphasized above.  The potential for
stratospheric ozone depletion can clearly be reduced by a given amount
using different technical options to restrict any particular POD in a
given year.   But it can also be done by reducing the production and use
of different PODs or of given PODs in different years.   We do not
attempt to examine such alternatives here in detail, but the framework
and data developed here can be applied to a wide range of questions
phrased as this form of cost-effectiveness analysis.

-------
                                - 59 -
                            Appendix B

         MODELS AND  DATA USED TO QUANTIFY TECHNICAL
                         CONTROL OPTIONS

     We used the methodology developed in Palmer et al. (1980),
Appendixes D and E,  to  estimate the prices at which foam producers and
solvent users voluntarily switch between production processes.   Palmer
et al. (1980) define that methodology  in detail.  This appendix provides
a brief review and  the  updated data used to calculate these prices.

FLEXIBLE SLABSTOCK FOAM
     Palmer et al.  (1980, pp. 53-58, 267-270) treated both slabstock and
molded flexible foams.  New technical  options have been developed
recently to displace CFC-11 in molded  foam production.  Unfortunately,
we do not have good data on this technical advance.  Hence, only the
data for slabstock  foams have been updated.
     In this analysis,  we develop a cost function to define the cost of
producing a given quantity  of foam under four different arrangements.
Case 1, the "base case," assumes that  half of the auxiliary blowing
agent is CFC-11 and the remainder is methylene  chloride.  It also
assumes that no recovery and recycle is used.   Interviews with industry
sources suggest that this approximates the typical circumstances in
industry today. Case 2 assumes that recovery and recycle equipment is
purchased and used.   Case 3 assumes no recycle  and recovery, but allows
the share of jnethylene  chloride to rise to 75 percent, about as high as
it can reasonably rise  according to  interviews  with industry sources.
Case 4 assumes that methylene chloride accounts for 75 percent of the
auxiliary blowing agent and that recycle and recovery  is under way.
     For Cases 1 and 3, the cost function  looks like the following:
     TC = (P  C + P  M) f + (P  A + P  M)(l  -  f) o       (B.I)
            cm          am

where P 'is the price of CFC-11, C is the quantity of CFC-11 demanded at

-------
                                  - 60  -
quantity of nonblowing materials demanded, f  is the share of CFC-11 as
an  auxiliary blowing  agent, P   is the price of the alternative
                             £1
(methylene chloride), A  is the  quantity of the alternative demanded at
full  capacity, and a  is  an adjustment factor  to show how the use of the
alternative affects productivity.
      For Cases 2 and  4,  the cost function looks like the following:
     TC =  [P  C(l-e)+bCe + P  M] f +  [P  A  (1 - e)
            C                      ID          fl.
          + b A
e+P  M](l - f)  and e.  Updated values of

-------
                 - 61 -


               Table B.I

DATA INPUTS FOR FLEXIBLE SLABSTOCK FOAM
Parameter
Annual plant use of CFC-11, if
100% CFC-11, Ib/yr, C
Share of market
1984 price of CFC-11, $/lb, P
Price of methylene chloride, $/lb, P
Ci
Annual cost of nonblowing agents,
Small
Plant
150,000
0.19
0.51
0.26
1,430,000
Medium
Plant
225,000
0.28
0.51
0.26
2,140,000
Large
Plant
1,200,000
0.53
0.51
0.26
1,140,0000
  $/yr, Pm M

Annual plant use of meth chlor if
  converted, Ib/yr, A

Capital costs of recovery
  equipment, $, K

Operating costs per Ib of CFC of
  recovery equipment, $/lb, b

Other annual operating costs of
  recovery equipment, $/yr, 0

Discount factor, X
  (10 yr life, r = 0.15)

Fraction of blowing agent
  recovered, e

Material cost weighting factor
  w/o conversion, a_

Material cost weighting factor
  with conversion, a
                 *  c
Fraction of foam blown with CFC
  w/o conversion, f.

Fraction of foam blown with CFC
  with conversion, f
                           0.85*C     0.85*C
                           0.50
                           1.01
                           1.07
                           0.50
                           0.25
0.50
1.01
1.023
0.50
0.25
            0.85*C
                         1,000,000  1,000,000   1,000,000
                           0.0175     0.0175      0.0175
                            95,000     95,000      95,000
                           0.19925    0.19925     0.19925
0.50
1.01
1.01
0.50
0.25

-------
                                 - 62 -
 P   and  P   are drawn  from the International Trade Commission and various
 C      a
 issues  of  Chemical Marketing Reporter.  The following items were revised
 on  the  basis of recent  interviews with industry to reflect updated views
 of  costs and cost structures:  P  M, b, and 0 .  The value of X reflects
                                m            r
 a value of the real  pretax cost of capital equal to 0.15 rather than
 0.20; a value of 0.15 is more in line with current costs and investment
 tax arrangements than 0.20.  The values of a have been revised down from
 1.125 on the basis of sensitivity analyses for specific types of plants.
 The sensitivity analyses sought to achieve a greater degree of
 correspondence between  the results of this cost analysis and observed
 patterns in industry.   The values of f reflect a shift toward use of
 methylene  chloride since Palmer et al. was written.

 EXTRUDED POLYSTYRENE SHEET
     The analysis of PS sheet is analogous to that for flexible foam
 above.  The analysis considers three arrangements whose total,
 annualized costs can be expressed using equations like those for
 flexible foam.  A minor change is required in the characterization of
 materials  used for PS sheet because its production with CFC-12 and
 pentane involves different amounts of materials.  We define M0 as the
                                                             \j
 quantity of materials relevant to CFC-12 and M. as the quantity of
                                              A
 materials  relevant to the alternative, pentane; we substitute these for
 M in the equations as appropriate.   Then the three arrangements can be
 defined as follows.   In Case 1, the base case, f = 1, so that no
 alternative gas, in this case pentane, is used.  Hence, Eq. (B.I)
 becomes

                           TC = PC C + Pm MC                   (B.3)

 In Case 2,  no alternative gas is used, but recovery and recycle
equipment  is added.   From Eq. (B.2), we find

     TC = Pc C (1 -  e) + b C e + PM Mc + Or + X Kr             (B.4)

Case 3 allows total  conversion to pentane to that recovery and recycle
 is no longer a concern.   Such conversion involved a capital investment

-------
                                 -  63 -

(K ) and new operating costs (0 ),  which can be related to  total,
annualized costs as follows:

                     TC=P  A+P   M.+O  +XK              (B.5)
                           a     m  A    a      a

With appropriate data, these cost functions can be expressed as linear
functions of the price of CFC-12 and switch prices calculated.   Table
B.2 provides the data used to quantify these equations.  These  data do
not reflect any new discussions in  the industry.  They simply update
data in Palmer et al. (pp. 112-119) for changes in prices and the cost
of capital.  The following values are the same as those in Palmer et
al.:  C, share of market, A, M_, M., e,  a, and f.  P , P ,  and  P  are
                              u   A                 cam
updated using information from the International Trade Commission and
Chemical Marketing Reporter.  K  is the same as the value used  in Table
B.I, and b is updated from 1976 to 1984, using the average hourly
earnings per worker  (nonseasonal) for chemicals and related products,
reported in the Survey of Current Business, to provide an index of
rising labor costs.  0  is updated from 1976 to 1984 using the producer
price index for fuels and related products and power, reported in the
Survey of Current Business.  X is revised downward to reflect a real
pretax cost of capital of 0.15.

SOLVENTS:   CFC-113
     The treatment of CFC-113 here is somewhat  different from that in
Palmer et al. (1980, pp.  78-84, 271-277).  The  frame of analysis--types
of drying and cleaning units--has changed and the technical options
examined has expanded.  The analysis now considers six unit types,
small, medium, and large  in size, and using  either pure CFC-113 or
CFC-113 in combination with something else.  Palmer et al.  (1980)
considered a finer grid of  distinctions by  size but could not  examine
the importance of the "pure" versus  "combination" distinction.3  The
analysis also considers carbon adsorption  for recovery of CFC-113 vapor
losses.
      3The unit sizes used here  can be  related to those  in Palmer et al.
Table E.I, p. 272, as  follows:  The  small unit here  is  Case 2, medium
unit  is Case 4,  and large unit  is Case 8.

-------
                                 - 64 -


                               Table B.2

             DATA INPUTS FOR THERMOFORMED POLYSTYRENE SHEET
Parameter
Annual plant use of CFC-12,
Ib/yr, C
Share of market
1984 price of CFC-12, $/lb, P
C
Price of pentane, $/lb, P
fl.
Price of nonblowing agent
Small
Plant
350,000
0.25
0.67
0.17
0.45
Medium
Plant
500,000
0.50
0.67
0.17
0.45
Large
Plant
750,000
0.25
0.67
0.17
0.45
  materials, S/lb, P
                    m
Annual plant use of pentane if
  totally converted, Ib/yr, A

Annual plant use of nonblowing
  materials with CFC-12, Ib/yr, M,,
                                 o
Annual plant use of nonblowing
  material with pentane, Ib/yr, M.
                                 A
Capital costs of recovery
  equipment, $, K

Operating costs per Ib of CFC
  of recovery equipment, $/lb, b

Other annual operating costs of
  recovery equipment, $/yr 0

Discount factor, X
  (10 yr life, r = 0.15)

Fraction of blowing agent
  recovered, e

Material cost weighting
  factor, a

Fraction of foam blown with
  CFC, f
  350,000    500,000     750,000
4,487,000  6,410,000   9,615,000
4,936,000  7,051,000  10,577,000
1,000,000  1,000,000   1,000,000
   0.0175     0.0175      0.0175
   95,000     95,000      95,000
   0.19925    0.19925     0.19925
   0.50
0.50
0.50

-------
                                 - 65 -
     Despite these changes, the basic methods used have not  changed
much.  For each type of unit, we examine seven alternatives  or  cases,
not all mutually exclusive:

     1.  Continuation of the status quo,
     2.  Use of more conservative equipment to reduce vapor  loss,
     3.  Use of recovery and recycle equipment to reduce vapor  loss,
     4.  Joint use of more conservative and recovery and recycle
         equipment,
     5.  Use.of in-house distillation to reduce waste loss,
     6.  Use of external reclamation services to reduce waste loss,
     7.  Switching to an alternative solvent.

The structure of our cost models, which appears to be reasonable,  allows
us to treat vapor and waste losses separately, as they were in Palmer et
al. (1980, pp. 271-277).  We construct cost functions like those above
to compare Cases 1 through 4 and then Cases 1 and 5.  External
reclamation and switching to an alternative solvent are treated
differently.  We consider reclamation and switching first and then we
will return to the other options.

Reclamation and  Switching
     We assume that there are important sources of heterogeneity in
cleaning and drying applications of CFC-113 that  are not captured in the
formal analysis of cost functions.  Some units are  located near external
reclamation services and hence  can take advantage of these services at a
lower cost than other units.  Similarly, some solvent  applications are
more amenable to the use of  an  alternative solvent  than others.  We
assume that as the price of  CFC-113 rises, more and more units will find
external reclamation and switching attractive on  cost  grounds.  Hence,
we assume that CFC-113 use will fall  at a certain rate as its price
rises, causing more owners of units to  reduce their waste losses through
external reclamation or to cut  their  use to  zero  by moving to  another
solvent.  Empirical data are not readily available  to  tell us how fast
this might occur.  As the  text  indicates, the rate  we  pick  is more

-------
                                  - 66  -
 important  for switching than  for reclamation because our analysis is
 more  sensitive to the choice  of the first rate.  Based on very rough
 subjective judgment, we assume that CFC-113 use falls 1 percent for
 every 1 percent rise in price as a result of movements to alternative
 solvents;  use falls one-half  a percent for every percent rise in price
 as more users find it worthwhile to ship their waste to a reclaimer.
 The effect of changes in price does not occur for switching until a
 certain threshold price is reached.  The threshold price represents the
 price increase required to cover the cost of new capital equipment that
 makes a switch possible.  We  use the same concepts as those used in
 Palmer et  al. (1980, p. 81) to define thresholds for specific units.  No
 investments are required for  external reclamation; hence, use reductions
 start immediately if the price of CFC-113 rises.  Cuts in the size of
 the CFC-113 market caused by  switching to alternative solvents and
 external reclamation reduce the number of units that can be affected by
 other actions, but not the costs at which those actions become cost-
 effective.

 Other Reduction Options
      Within the market for CFC-113 that remains after switching and
 external reclamation have occurred, the other actions that can be taken
 continue to be taken.  The prices at which'these actions become cost-
 effective  can be calculated using the same kinds of cost functions used
 for the foams above.  The primary differences are that uses associated
with  replacing vapor and waste losses must be carefully distinguished
 and material costs no longer  play a role.
      For the first, base case, then, total, annual cost can be
 represented by an equation much like Eq. (B.3):
                       TC = Pc VQ + (Pc + Cw) WQ               (B.6)

where V. is initial vapor loss per year, W. is initial waste loss per
year, and C  is the cost of disposing of waste.  Case 2 requires an
investment (K ) and new operating costs (OE ) that reduce vapor loss by
             C                             C
a factor 6:

-------
                                 -  67  -
     TC = Pc VQ 6 + Xc Kc + OEc + (Pc + CJ  WQ                 (B.7)

Similarly, using carbon adsorption equipment to recover and recycle
solvent cuts vapor losses by a factor (1 - &), but requires new
investment (K ) and operating costs (OEfl), as well as the cost of
reconstituting a combined solvent following its recovery (ffl):

                    TC = Pc VQ (1 - « + Xfl Kfl + OEfl

                         +  (Pc + V W0 + fa                    (B'8)

Combining these two options, in Case 4,  leads to joint effects on vapor
losses and all of the new costs of both:
     TC = Pc VQ fl  (1 - P) + Xc Kc + Xa Kfl + OEc + OEa.

          +  (Pc +  Cw) W0 + fa                                   (B'9)

Case 5, using  in-house distillation equipment, has  full  annual  costs
that look much like those for carbon  adsorption  and for  similar reasons.
Vapor  losses fall  by a factor  (1  - Y) at the  cost of new investment
(K,),  operating expenses  (OE .) , and reconstitution  of .combined  solvent
following distillation  (f,):
                         a

                    TC =Pc VQ  +WQ  (Pc+Cw)(l  -  Y)

                         +  X K . + OE , + f .                      (B.10)
                               a     da.

     Each of these cost  functions is  linear in the price of CFC-113.  By
substituting the  relevant  quantitative  data,  we can use these functions
to calculate price ranges  over which each  of these options is cost-
effective.   These prices are  reported in the text.
     The data used to quantify these cost  functions are shown in Table
B.3.   They  are based in part  on data from Palmer et al.  (1980)  and  in
part on new data from industry.  The following variables are treated

-------
                                   - 68 -
                                Table  B.3

                         DATA INPUTS  OF CFC-113
Parameter
Market share
Annual vapor

loss,
Snail
Pure
.24
1931
Plant
Cofflbin
.24
1931
Medium Plant
Pure Comb in
.IS
6610
.15
6610
Large Plant
Pure Comb in
.11 .11
30670 30670
 lb/yr,

 Annual waste  loss,
Waste disposal cost,
Vapor  loss saving in
  cons, equip., 6

Reuse  efficiency for
  carb. adsorb., B

Efficiency of in-house
                             971    971     1869    1869      4369   4369
                             538    579     1217    1297      2846   3032
                            .57    .57      .49     .49       .14     .14
                                                     -same-
                                                     -sane-
Efficiency of external
  reel., Ie, r.

Discount factor, cons.
  equip., X
                            .95
                            .22285
                                                     •saoe-
Discount factor,  carbon     .19925  ---—	----same—	
  adsorb.,  X

Discount factor,  in-house    .22285	same	———
  distilling, X.
               a
Capital cost for conserv.     9530  9530    13500 13500  224800 224800
  equip., $/unit KC

Capital cost for carbon
  adsorp. $/unit, K

Capital equip for in-
  house dist., $/unit, K.
                        a
Operating cost for cons.
  equip., OE,.
                             7000  7000    21000  21000    87500  87500
                             3930  3930     9170   9170    13100  13100
                                                     -sane-
Operating cost for carbon   .05*K
  adsorp., OE

Operating cost for in-         0
  house dist., OE.
                 a

Added cost for carbon
  adsorp., S, f^

Added cost for in-houst
                                                     -sane-
                               0   9340
                               0   2670
                                                   9340
2670
0     9340


0     2670
Transport cost for ext.         0
  reclan, $/lb, a
                                                     -sane-

-------
                                 - 69 -
just as Palmer et al. treated them:  market share, Vn, Wn, 6, Y, OE ,
                                                    u   u          o
OE,, and a.  C  Wo is based on Palmer et al., but updated to reflect a
disposal cost of $300 a barrel.  |5, X ,  K , OE , and f  are relevant to
                                     a   a    a       a
the new option of carbon absorption of vapor losses.  Their values are
based on information in Mooz et al. (1982).  K  and f  are updated from
                                              a      a
1980 to 1984 with price indexes.  K  is updated using the Bureau of
                                   d
Labor Statistics PPI 1166.04, producer price index for chemical industry
machinery.  F  is updated using average hourly earnings per worker
             &
(nonseasonal) for chemicals and related products reported in the Survey
of Current Business.  X  and X, are updated to reflect a real pretax
cost of capital of 0.15.  K  uses data from Palmer et al., updated from
1976 to 1984 using BLS PPI 1166.04.  Kd uses data from Mooz et  al.,
updated from 1980 to 1984 using BLS PPI 1166.04.

-------
                                 - 70 -


                             Appendix C

      DIFFICULTIES IN USING U.S.  RESULTS TO  LOOK ABROAD

     The results reported here draw heavily on information about
technical options that use technologies available not  only in the United
States, but in most other parts of the world.   Similarly,  the United
States has no special claim on the technologies that underlie
alternatives that allow product substitution.   Aner:innovation that
generates new technological opportunities in the United States or
elsewhere should similarly benefit all parts of the world  with access to
this new knowledge.  These considerations suggest that the kind of
information developed in this Note for the United States could easily be
transferred to other areas as well.  Unfortunately, the problem is not
so simple.  Differences in prices, regulations, and markets can make
some options that work in the United States irrelevant in  other parts of
the world.  Similarly, options not mentioned here may  be extremely
important elsewhere.  This appendix discusses  these differences briefly
as a form of caveat to the reader tempted to use results reported here
outside the appropriate context of the United  States.

DIFFERENCES  IN PRICES AND THE COST  OF CAPITAL
     Prices of PODs, the products or processes in which they are used,
                          \,
and cofactors used to make jthose products are  all important to
opportunities to reduce the use of these PODs.
     Consider product substitution.  In the United States, the prices of
packaging materials based on CFC-12 and wood fiber are so  close in many
parts of the United States that demand for CFC-12-based products is
highly sensitive to price.  This  need not be the case  in other parts of
the world where wood is more or less  expensive.  Because wood is a lower
value commodity per pound than CFC-12 or plastic resin, its price is
likely to be more sensitive to transportation  costs and hence location.

-------
                                 - 71 -
for CFC-11, -12, and -113.  Although these substances are traded in a
global market, suggesting that prices in one country must at least  be
related to prices in another for any chemical,  transportation costs,
duties, taxes, and regulations can lead to significant differences. We
have not attempted to measure these differences because our study is
limited to the United States.  But we would expect to find them if we
look more closely at other countries.
     Differences in the costs (and productivity) of cofactors like labor
are much better known.  These costs are quite important to certain
aspects of reducing the use of CFC-113 and recovering CFC-12 from mobile
air conditioners.  They also play a role in foam making.  They could
potentially be important to options not considered here, primarily
because the high cost of U.S. labor makes them too costly.  An example
is removal of rigid foam insulation from buildings.
     One of the most important differences may be in the cost of capital
in different countries.  Recall that decisions to recover and recycle,
to buy more conservative equipment, to use in-house distillation,  and
often even to switch to a substitute chemical depend on  investments,
whose costs must be justified by future savings of operating costs.  The
cost of capital plays a vital role in determining whether such
investments are worthwhile.  Some might argue that in the current  world
of international finance, all industrial borrowers face  the same costs
of capital around the world.  Local investment climate—what
international investors often speak of in terms of "country risk"--
easily creates differences  from one country to another.  But even
setting aside this sometimes intangible source of differences,  local
taxes and duties routinely  create large differences  in the cost of
capital from one country to another.  For example, King  and Fullerton
(1984) reports that effective tax rates on investment in machinery in
Sweden, the United Kingdom,  the United States, and West  Germany are
respectively 0.2, -36.8,  17.6, and 44.5 percent.  Among  four nations
typically thought of as being heavily industrialized market economies
and comparable in technology, POD users in the United Kingdom would
invest to reduce their use  of PODs at much  lower prices  than our results
would suggest--and may have already--whereas users in West Germany would
require much higher prices  than those in the United  States.  The

-------
                                 - 72 -
diversity of tax rates, with their implications for the cost of capital
and willingness to invest, is likely to be still greater as we look at
more diverse countries.

DIFFERENCES  IN REGULATION
     The substances we suggest as substitutes for PODs have a number of
negative properties that may lead other countries to regulate them
differently from the way the United States regulates them.
     For example, some governments are much more concerned about fire
hazards than U.S. officials typically are.  The Japanese prohibit the
use of hydrocarbon propellants in aerosols because of the additional
fire hazard they create; this may suggest that they would not look
favorably on using pentane in foam blowing.  At the very least, they are
likely to require more costly safeguards in its use as a blowing agent
than those required in the United States.  West Europeans also appear to
be more stringent about the location of aerosol canning plants that use
hydrocarbons; this concern could also be reflected in their treatment of
pentane as a foam blowing agent.
     West Europeans also show greater sensitivity to the possible
carcinogenicity of methylene chloride than U.S. officials do.  This
leads to a marked difference in foam blowing practices between the
United States and Europe.  European foam blowers typically do not use
methylene chloride despite its cost competitiveness with CFC-11, in part
because it may pose a threat to health.  This would limit the
applicability of this substitution option in Europe or at least
significantly raise its perceived social cost.
     Similar problems are likely to arise with the use of CFC-113 and
methyl chloroform.  A major opportunity for reducing their use in the
United States is to substitute toward other chemicals, many of them
chlorinated solvents that may present serious health hazards and
disposal problems.  Although we have no direct evidence of it, we would
expect other countries to view those problems differently.   Indirect
evidence comes  from the alleged widespread use of carbon tetrachloride
as a solvent in the Third World, evidence of lower concerns about health
effects in this part of the world.

-------
                                 -  73  -
OTHER DIFFERENCES
     Differences in use and production patterns  are likely  to  lead to
large differences in the relative importance of  options  considered here
and in the effectiveness of any one of them in reducing  potential risks
to stratospheric ozone.  The most obvious case is the absence  of any
discussion of aerosols in this Note.  Reduced use of CFCs  in aerosols
probably provides the least costly way to reduce substantially the
threat to the ozone layer from emissions outside the United States.
Widespread use of carbon tetrachloride as a solvent outside the United
States is another example; it deserves attention that would not be
appropriate here.  More generally, mobile air conditioning is  likely  to
be less important outside the United States.  Air conditioning in
general is not important in parts of Europe that are heavy users of
CFCs.  In sum, these differences in use and production patterns make  it
clear that the effectiveness of individual measures will differ
significantly outside the United States.  And the summary results  of
this study should not be transferred elsewhere.
     Although the technologies that underlie the choices discussed here
would probably be usable anywhere in the world,  then, the decisions to
use them remain behavioral, not technological.  And the factors that
affect behavior--prices and regulations--can differ substantially from
one country to another.  At the very least, readers should  exercise
caution in using information about  the incremental social cost of a
particular decision in the United States in a different setting.  More
generally, the patterns of use and  production in the United States are
likely to differ profoundly from those elsewhere.  Conclusions about the
effectiveness of individual decisions or about  the cost and
effectiveness of a program  like one undertaken  in  the United  States to
reduce the use of PODs are highly unlikely  to be useful in  another
context.

-------
                                 -  75  -


                             REFERENCES
E. I. DuPont de Nemours and Company,  Microfoam Polypropylene Foam:  A
  Product Made Using Chlorofluorocarbon Blowing Agents,  submission to
  the U.S. Environmental Protection Agency,  1978.

Hammitt, J. K., et al., Product Uses and Market Trends for Potential
  Ozone-Depleting Substances:  1985-2000, The Rand Corporation,
  R-3386-EPA, May 1986.

Harberger, A. C., "Three Basic Postulates for Applied Welfare
  Economics," Journal of Economic Literature, Vol. 9, pp.  785-797,
  September 1971.

King, M. A., and D. Fullerton (eds.), The Taxation of Income from
  Capital, University of Chicago Press, Chicago, 1984.

Mooz, W. E., K. A. Wolf, and F. Camm, Potential Constraints on
  Cumulative Global Production of Chlorofluorocarbons, The Rand
  Corporation, R-3400-EPA, May 1986.

Mooz, W. E. et al., Technical Options for Reducing Chlorofluorocarbon
  Emissions, The Rand Corporation, R-2879-EPA,  March 1982.

National Academy of Sciences, Halocarbons:  Effects of Stratospheric
  Ozone, Washington, D.C., 1976.

National Academy of Sciences, Protection against Depletion of
  Stratospheric Ozone by Chlorofluorocarbons, Washington, D.C., 1979.

National Academy of Sciences, Causes and Effects of Stratospheric Ozone
  Reduction:  An Update, Washington, D.C., 1982.

National Academy of Sciences, Causes and Effects of Changes in
  Stratospheric Ozone:  Update 1983, Washington, D.C.,  1984.

Palmer, A. R., W. E. Mooz, T. H. Quinn,  and K. A. Wolf, Economic
  Implications of Regulating Chlorofluorocarbon Emissions from
  Nonaerosol Applications, The Rand Corporation, R-2524-EPA,  June 1980.

Palmer, A. R. and T. H. Quinn, Economic Impact Assessment of a
  Chlorofluorocarbon Production Cap, The  Rand Corporation, N-1656-EPA,
  February 1981.

Ramanthan, V., R. J. Cicerone, H.  B. Singh,  and J. T. Kiehl,  "Trace Gas
  Trends  and Their Potential Role  in Climate Change," Journal of
  Geophysical Research, Vol. 90, p.  5547, 20 June  1985.

-------