Jump to ContentJump to Main Navigation
Protecting the Ozone LayerScience and Strategy$

Edward A. Parson

Print publication date: 2003

Print ISBN-13: 9780195155495

Published to Oxford Scholarship Online: November 2003

DOI: 10.1093/0195155491.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2020. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use.  Subscriber: null; date: 27 February 2020

Industry Strategy and Technical Innovation, 1987–1992

Industry Strategy and Technical Innovation, 1987–1992

(p.173) 7 Industry Strategy and Technical Innovation, 1987–1992
Protecting the Ozone Layer

Edward A. Parson (Contributor Webpage)

Oxford University Press

Abstract and Keywords

Examines the major changes in technology that followed the adoption of the Montreal Protocol in 1987 and the completion of the period of initial formation of the ozone protection regime. Examines the reactions of major producers and users of chlorofluorohydrocarbons (CFCs) to the challenge posed by the negotiation and adoption of the Protocol. Discusses how major CFC producers in the chemical industry revived previously abandoned efforts to commercialize less ozone‐damaging chemical alternatives to CFCs. Also examines how CFC users, many of whom faced more serious risks from CFC restrictions than the producers, responded with intense efforts to reduce their dependence on all ozone‐depleting chemicals — efforts that reduced most CFC uses much faster than had seemed possible, and directed large shares of former CFC markets away from similar chemicals entirely.

Keywords:   alternatives, CFCs, chemical industries, chlorofluorohydrocarbons, environmental protection, international agreements, markets, Montreal Protocol, ozone depletion

7.1 Chlorofluorocarbon Markets After the Protocol

The 1987 Protocol exerted immediate pressure on industries producing and using CFCs worldwide, which grew more intense over the following two years as new scientific results and assessments led to a growing chorus of calls to eliminate CFCs. Firms responded with a flood of innovation to reduce use of CFCs and other ozone‐depleting chemicals, which was substantially faster and more diverse than even the strongest advocates of controls had anticipated. At the time of the Protocol, CFC markets reflected trends that had persisted since the late 1970s. Driven by non‐aerosol uses, world market growth of CFCs 11 and 12 continued after resuming in 1983, and passed 1 million tonnes in 1988.1 Despite the resumption of growth, these remained low‐margin commodity markets throughout the industrialized countries. The mix of uses differed substantially between the United States and the rest of the world. Aerosols were a smaller share of U.S. CFC use even before their drop in the 1970s, and by 1986 represented only a few percent of use in the United States and other countries that had banned aerosols, but 30 percent in the rest of the world and 50 percent in Europe. All other uses were proportionately larger in the United States, with an additional large increment for automobile air‐conditioning.2 Halon markets faced similar problems of soft demand and low margins in the 1980s, with prices falling 30 percent between 1980 and 1987. The only strong CFC businesses were CFC‐113 and sales to developing countries, where markets grew rapidly through the 1970s and 1980s. While some CFC 11 and 12 were produced in developing countries, most of this growth was served by European producers, who dominated world exports.3 Worldwide, about one‐quarter of CFC use was in refrigeration (of which 40 percent was automotive air‐conditioning), one‐quarter in foams, one‐quarter in aerosols, and one‐quarter in solvents and miscellaneous uses. Contrary to popular opinion, home refrigerators were a tiny use, only about 1 percent.4 (p.174)

                   Industry Strategy and Technical Innovation, 1987–1992

Figure 7.1 World Production and Consumption Of Ozone‐Depleting Chemicals. In 1986, CFCs 11 and 12 Still Represented More Than Two‐Thirds Of World Production Of Ozone‐Depleting Chemicals (With Chemicals Weighted By Ozone‐Depletion Potential), Although the Contributions Of CFC‐113, Methyl Chloroform, and Halons Were Growing Rapidly. Aerosols Were Only a Few Percent Of Consumption In the United States and a Few Other Countries, But More Than 25 Percent In the Rest Of the World. Source: Author's Calculations, Based On Data From AFEAS 1995; Midgley 1989; EPA 1988. UNEP 1999b; UNEP 2002.

The Protocol began to provoke large‐scale reorganization of CFC industries and their political representation almost immediately. In the United States, the two smallest producers (Kaiser and Racon) announced their intention to leave the market.5 At the same time, the Alliance began to assume a more prominent role as the voice of U.S. industry in international discussions, gradually eclipsing the CMA Panel, and the Alliance's executive director increasingly acted as spokesman for the organization, rather than the DuPont official who served as science adviser.

In Europe, the most immediate effect was a rapid drop in aerosol use, spurred by the consumer boycott initiated in August 1987, which reduced aerosols to 19 percent of world CFC use by 1988.6 For nonaerosol sectors, however, there was no similar basis for confidence that CFC use could be easily and rapidly reduced. While the Protocol has been widely praised as precautionary because it was adopted under great uncertainty about the severity and immediacy of the ozone‐depletion risk, it was even more striking for the extreme technological uncertainty under which it was adopted regarding the cost and feasibility of cutting nonaerosol CFC uses. In September 1987, no one knew what the technical challenges and costs would be of cutting CFCs by half in the nations that had already eliminated aerosols.7

(p.175) The stakes were high, because many of the goods and services provided using CFCs, most notably refrigeration and electronics, were essential. Although CFCs were a small market, only about $2 billion worldwide, this quantity substantially understated their contribution to the world economy by an amount that was probably large, highly uncertain, and intensely contested. As industry argued with some justification, CFCs were intermediate goods that were incorporated into other products of substantially higher value that depended on them. Although the value share of CFCs in many final products was small, ranging from less than 0.1 percent in refrigeration and electronics to about 20 percent in insulating foams, the precise technical properties of CFCs were in many cases essential to the performance of the final products. Moreover, a great deal of capital equipment was optimized to, or dependent on, the precise properties of a particular chemical. In 1988, $28 billion of goods and services were produced with CFCs in the United States, using $135 billion worth of capital equipment ($385 billion worldwide).8 CFCs were supplied competitively, with substantial overcapacity in the industry, so their producers could not capture more of the chemicals' contribution through raising prices. But potential limits in technical substitutability meant that the cost of cutting CFCs could substantially exceed their price, and in the extreme could approach the premature loss of all this capital. At this time, very little was known about the actual ease or difficulty of replacing CFCs. Some uses appeared to have very limited short‐run substitution possibilities, but the experience of the 1970s aerosol cuts suggested that substitution difficulty might be overestimated in advance, while the extreme diversity of uses suggested that at least some should be easy to replace.

7.2 The Pursuit of Chemical Substitutes

The major CFC producers revived their programs to develop chemical alternatives to CFCs in 1986 and early 1987. Three promising candidates were already in commercial production (HCFC‐22 at large scale, HCFC‐142b and HFC‐152a at small scale), although none of these was an ideal substitute for either CFC 11 or 12.9 The substitutes whose thermodynamic properties made them the most promising—HCFCs 123 and 141b for CFC 11, HFC‐134a for CFC 12—all required pilot‐scale production for toxicity and process testing, and development of a commercial synthesis route, which were estimated to take about five years in total.10 Early estimates suggested that the cost of HCFC‐123 could be $1.50 to $2.00 per pound, and that of HFC‐134a around $3.00, compared with about 65 cents per pound for CFC‐12.11 All proposed substitutes also needed varying degrees of development to make them work in existing applications, although an independent expert group convened by the EPA in late 1986 had concluded that this could be readily achieved under regulatory conditions that made the higher‐priced substitutes viable.12 The signature of the Protocol and the widely perceived risk of tighter restrictions to follow gave additional impetus to efforts to develop alternatives.

An important early step in developing alternatives was the formation of cooperative bodies to conduct toxicity and environmental testing of the most promising candidates. Since most proposed substitutes had not previously been marketed in commercial volumes, they required slow and costly testing for toxicity and environmental acceptability. Despite the intense competition under way to commercialize substitutes, it was easy to cooperate on these programs because the knowledge gained could not be withheld or used for commercial advantage. Thirteen CFC (p.176) producers from the United States, Europe, and Japan announced the formation of a cooperative program for toxicity testing, the Program for Alternative Fluorocarbon Toxicity Testing (PAFT), in January 1988.13 PAFT received authorization from antitrust officials in the United States, Europe, and Japan. In the United States, such protection was provided by the National Cooperative Research Act, which allowed competitors to collaborate on research serving the public interest. Testing under PAFT began in early 1988 with a two‐year inhalation study of HCFC‐123 and HFC‐134a.14 Two additional studies, each supported by a slightly different group of firms, tested HCFC‐141b (a substitute for CFC‐11 in foams) and HCFC‐124 and HFC‐125 (substitutes for CFCs 114 and 115 in refrigeration and sterilization).15 In December 1988, 12 producers formed a similar cooperative group to test alternatives for environmental effects, the Alternative Fluorocarbon Environmental Acceptability Study (AFEAS). In contrast to toxicity testing, for which procedures were well established, environmental screening of new chemicals was (and is) a less well‐specified task, so AFEAS had to develop criteria for environmental acceptability in a rapidly developing regulatory environment. After a small initial study, it was decided to assess each chemical by its contribution to ozone depletion and global warming, and the fate of its tropospheric breakdown products. AFEAS studied 12 chemicals—the same five as PAFT, plus three others that were already marketed and whose toxicology was known (HCFCs 22 and 142b, and HFC‐152a) and four that were initially regarded as second‐rank candidates (HCFCs 225ca and 225cb; HFCs 32 and 143a). AFEAS also took over the job of reporting CFC production when CMA's Fluorocarbon Panel ceased operation in 1990, and gradually expanded the scope of reporting to include HCFCs.16

While these new HCFCs and HFCs saw the most intensive development effort, there was initially also substantial interest in HCFC‐22 as a CFC substitute. Its situation was unique, since it was a poor match for the thermodynamic properties of CFCs 11 and 12, but was already manufactured cheaply at large volume and its capacity could be expanded rapidly by converting existing plants. Although producers highlighted their efforts to develop new chemicals, several also initially undertook large expansions of their HCFC‐22 capacity and promoted it as a near‐term substitute, especially for aerosols and foam food packaging.17

The largest prize in this immediate competition appeared to be the European aerosol market. Through 1987, European producers rapidly expanded their HCFC‐22 capacity, while DuPont conducted a major European sales drive for both HCFC‐22 and nonfluorocarbon propellants.18 In February 1988, however, ICI shocked its competitors by announcing it would not market HCFC‐22 for toiletries, the bulk of aerosol markets, because of lingering concerns about its toxicity. Other producers were highly critical of ICI's decision, including DuPont—which had refrained from promoting 22 as a substitute for U.S. aerosols 10 years earlier out of similar concerns but had since decided it was safe enough. Still, all producers felt compelled to follow ICI's decision, eliminating European aerosols as a potential HCFC‐22 market.19

This market became rapidly less attractive in any case, as Friends of the Earth had increasing success with its boycott of aerosol products. As had happened 10 years earlier in the United States, aerosol‐product marketers could not hold a united front against this pressure in the face of their highly unequal dependence on CFC propellants. Two firms broke ranks in January 1988, and began labeling their non‐CFC products “ozone friendly.”20 The next month, eight firms (representing 65 (p.177) percent of the U.K. aerosols market) announced they would eliminate CFCs by the end of 1989. After an unsuccessful counterattack, the entire British aerosol industry followed in May 1988, agreeing to label non‐CFC products immediately and cut CFC use 90 percent by the end of 1989. The aerosol industry in the rest of Europe followed one year later, in an April 1989 agreement with the EC Commission to cut CFC use 90 percent (a 40 percent reduction in total European CFC use) by the end of 1990. After 10 years of claims to the contrary, it was suddenly clear that CFC aerosol propellants were as easy to eliminate, and as politically vulnerable, in Europe as they had been in the United States.21

Efforts to promote HCFC‐22 as a near‐term substitute in foam food packaging had more success. In the United States, HCFC‐22 received accelerated FDA approval in early 1988, first for fast foods and later for grocery display packaging, and was rapidly adopted in both markets.22 Driven by this and other new markets, world production of HCFC‐22 increased 50 percent between 1986 and 1992, then stabilized and began to decline. The decline reflected both specific limitations of 22 and a broader shift toward regarding all HCFCs as transitional substances, to be replaced in turn by nonozone‐depleting options as they become available.

The more important substitutes were the small set of one‐ and two‐carbon HCFCs and HFCs that had been identified in the 1970s. Although there are a few dozen such chemicals in total, a simple screening process based on each chemical's fractions of hydrogen, chlorine, and fluorine can identify those likely to meet required standards for nontoxicity, noninflammability, and short atmospheric life.23 The few chemicals that pass this screen are roughly the same set that was deemed serious enough to be studied by AFEAS and PAFT.

In addition to being acceptable on environmental, health, and safety grounds, desirable substitutes must have thermodynamic properties that make them suitable to replace CFCs in specific applications, and an economical synthesis route. The major CFC producers raced to develop these few obvious candidates with crash programs that many participants described as the most intense competition of their careers.24

For CFC‐11, different alternatives appeared most promising for its two major nonaerosol uses, blowing insulating foams and as the refrigerant in the large chillers of commercial air‐conditioning systems. Despite its moderate inflammability, the most promising alternative for foams initially appeared to be HCFC‐141b, and several firms quickly began to build or convert plants to produce it.25 DuPont announced a patent in October 1988 for a new process to coproduce HCFCs 141b and 142b in a single step, and announced plans to convert a plant to the new process.26 But long‐term growth prospects for HCFC‐141b were sharply curtailed in late 1988 when new model results suggested that its ODP was not 3 percent, as believed, but 10 to 12 percent, more than double that of other HCFCs.27 As these reports were verified, flexible foam makers abandoned their plans to switch to HCFC‐141b and producers began scaling back production plans.28 Although the EPA had initially encouraged switching to HCFC‐141b, it reversed its stance in 1993, banning its use as a solvent immediately and requiring that its production be phased out by 2003, nearly 30 years before most HCFCs.29

For refrigeration uses of CFC‐11, the most promising alternative appeared to be HCFC‐123. DuPont patented the first commercial synthesis route in mid‐1988, and announced a large plant using the new process in June 1989, planned for completion in late 1990 in Canada. HCFC‐123 production grew rapidly following its (p.178)

Table 7.1 The Major Ozone‐Depleting Chemicals, Their Uses, and Alternatives

Name and Formula

ODP (Steady‐state)

World Production (Kte, unweighted)

Major Uses

Major Alternatives (partial list)

CFC‐11 (CFCl3)


407 (1986)

Foams ∼ 60%

HCFCs 22, 123, and 141b; methylene chloride; hydrocarbons; HFCs 245fa and 365mfc; product reformulation

54 (1996)

Aerosols ∼ 30% (< 5% in US)

Ref/AC ∼ 8%

CFC‐12 (CF2Cl2)


463 (1986)

Auto AC ∼ 20% (>40% in US)

HFC‐134a, HFC‐152a, hydrocarbons, ammonia, CO2, aerosol product reformulation

119 (1996)

Other Ref/AC ∼ 25%

Aerosols ∼ 35% (<5% in US)

Foams ∼ 13%

CFC‐113 (C2F3Cl3)


197 (1986)

Solvents ∼ 100%

Aqueous and semi‐aqueous solvents, trichloroethylene, perchloroethylene, no‐clean processes, HCFC‐225

6 (1996)

CFC‐114 (C2F4Cl2)


19 (1986)

Foams ∼ 77%

HCFCs 22, 123, 141b; hydrocarbons; HFC blends; ammonia; CO2

0.7 (1996)

Ref/AC ∼ 23%

CFC‐115 (C2F5Cl)


12 (1986)

Ref/AC ∼ 100%

HCFC‐22, HFC blends, ammonia, hydrocarbons, CO2, water


Halon‐1211 (CF2BrCl)


15.5 (1986)

Fire fighting ∼ 100%

Operational changes to reduce emissions, management of existing stock for critical uses, water, dry chemical, CO2, inert gases

17.8 (1997)

(portable extinguishers)

Halon‐1301 (CF3Br)


11.2 (1986)

Fire fighting ∼ 100%

1.6 (1997)

(total‐flooding systems)

Methyl Chloroform (C2H3Cl3)


707 (1989)

Solvents ∼ 100%

Aqueous and semi‐aqueous solvents, trichloroethylene, perchloroethylene, dichloromethane, hydrocarbons, isopropyl alcohol


Carbon Tetrachloride (CCl4)


∼ 26 (1992)

Solvents ∼ 100%

< ∼ 5 (1996)

(ex >95% feedstock use)

Methyl Bromide (CH3Br)


46 (1986)

Fumigation: soils ∼ 75%

Heat, cold, solarization, chlorpicrin, 1,3‐D, metam sodium, dazomet, phosphine, sulfuryl fluoride, integrated pest management

(0.4 proposed)

69 (1996)

Structures, durables ∼ 16%

(ex <10% feedstock use)

Perishable goods ∼ 9%

HCFC‐22 (CHF2Cl)


268 (1989)

Ref/AC ∼ 85%

HFC blends

470 (1996)

HCFC‐141b (CH3CFCl2)


0 (1986)

Foams ∼ 88%

HFC‐152a, HFC‐134a, CO2, hydrocarbons, HFCs 245fa and 365mfc

121 (1996)

Solvents ∼ 9%

HCFC‐142b (CH3CF2Cl)−


7 (1986)

Foams ∼ 92%

HFC‐152a, HFC‐134a, CO2, hydrocarbons, HFCs 245fa and 365mfc

38 (1996)

(p.179) (p.180) introduction in January 1991, despite some continuing toxicity concerns and a 1989 suggestion that its atmospheric breakdown products could be environmentally harmful.30

To replace CFC‐12 the clear leading contender was HFC‐134a, a zero‐ODP chemical that matched CFC‐12's performance characteristics closely.31 HFC‐134a was the largest prize in the race for new chemicals because it was expected to serve the large automobile air‐conditioning market, and several firms pursued it via multiple synthesis routes. ICI made an aggressive commitment to HFC‐134a in early 1988 and opened their first commercial plant in October 1990, two months before DuPont.32 Through aggressive early development and marketing, as well as engineering work to help automakers solve problems of incompatibility with seals and lubricants, ICI claimed a substantial fraction of this market in the United States as well as in Europe.33 By 1991, DuPont and ICI each planned to have 134a plants operating in Europe, North America, and Asia by 1995. Allied, Atochem, Hoechst, Montefluous, Showa Denko, and Daikin all had plants under construction, while several other firms had announced building plans.34 Both DuPont and ICI announced important catalyst breakthroughs in 1992, which roughly doubled their capacity.35 Although rapid demand growth was projected, reaching 100 to 200 million pounds per year by the late 1990s, this building boom created an initial surplus in the face of softer than expected demand for the first few years.36

Despite intense development efforts, substitute chemicals faced several delays in coming to market at large scale, and often were not available in large enough quantities to support the transition of major uses until long after their availability was announced. As late as mid‐1991, the only alternatives commercially available were the three that were marketed before the Protocol, HCFCs 22 and 142b, and HFC‐152a. Several others were very close to commercial availability (e.g., HFC‐134a) or were awaiting results of toxicity and environmental testing (e.g., HFCs 32 and 125, HCFCs 123 and 141b).37 By late 1991, it was also becoming widely accepted that HCFCs would be used only as transitional substances. Attention increasingly shifted to HFCs, particularly to chemical blends that could match CFCs' properties better than any single alternative. HFC‐32 emerged as an especially important component in refrigerant blends, some approaching drop‐in performance, although its inflammability and high compressor discharge temperature made it problematic to use alone. Both DuPont and ICI opened HFC‐32 plants in the summer of 1992,38 and by 1993, DuPont, Allied, ICI, and Atochem were all marketing various patented refrigerant blends.39 After a few years of offering distinct, patented, but similarly performing blends that competed for the same applications, producers began cross‐licensing each other's blends that had come to dominate particular applications by the mid‐1990s. As production of chemical substitutes increased, manufacturers began consolidating and closing their CFC capacity in 1990.40

The race to develop and market these substitute chemicals posed large technical and commercial risks. Nearly all producers committed to ceasing production of CFCs before they had alternatives in hand, and committed to commercial‐scale production of promising alternatives before receiving final toxicity results or completing all process and application research.41 Although the industry cooperatives for testing toxicity and environmental impacts pooled some of these risks, the uncertainties associated with developing cost‐effective, large‐scale synthesis processes for new chemicals, and demonstrating their suitability in specific applications, represented large remaining risks. The producers that lacked the resources to pursue all promising chemicals simultaneously had to make large gambles on the success (p.181) of particular ones. This competition favored the largest producers and those with the strongest research capability, as the pattern of smaller producers being sold or closed within a few years demonstrated. The three smallest U.S. producers—including Pennwalt, which tried to compete in new refrigerant and foam markets—were all sold by 1989.42 Atochem's acquisition of Pennwalt and Racon, as well as ICI's construction of new U.S. plants, brought European producers a substantial fraction of U.S. production capacity for the first time, reaching about 25 percent by 1993.43

But even the largest producers faced serious risks in this transition. These were not limited to the technical risks inherent in learning how to produce the new chemicals profitably and make them work in existing applications, but included diverse market, regulatory, and political risks. For most applications a few different alternative chemicals were proposed, which competed on performance, ease of substitution, cost, and health and environmental impacts. The eventual cross‐licensing of refrigerant blends to eliminate proliferation of similar products reduced one such competitive risk, but others were more subtle and systematic. For example, in some applications near‐term HCFC substitutes (which were relatively cheap and easy to substitute in existing equipment) competed with HFCs or other longer‐term substitutes. Where expensive capital equipment such as refrigeration systems had to be designed around a chemical's properties, this competition among chemicals was mirrored in competition among equipment designs.44 Producers also faced the risk that chemicals they expended large efforts to develop would later be judged environmentally unacceptable or too toxic for the proposed application, as happened to HCFCs 22 and 141b.

The most acute risks were regulatory. Just as creating viable markets for higher‐cost alternatives depended on the incentives created by restriction of CFCs, so details of the regulatory treatment of CFCs and substitutes would greatly affect the profitability of all substitute markets, and could confer large advantages on one proposed class of substitutes or another. Weak or inadequately enforced CFC controls would gravely harm the viability of all substitutes, while too rapid a clamp‐down on CFCs could prevent an orderly transition, drive users into inferior alternatives that were immediately available, or impose large costs of premature capital retirement.45 Regulatory treatment of the new alternatives also posed serious risks, particularly in the United States, where the Significant New Alternatives Program (SNAP) under the 1990 Clean Air Act amendments gave the EPA authority to rule proposed CFC alternatives acceptable or unacceptable for specific applications, based on a broad set of environmental and safety criteria. Industry was harshly critical of this provision, especially EPA's refusal to promise that an alternative, once judged acceptable, would not subsequently be ruled unacceptable, arguing that the resultant uncertainty deterred the investments that were needed to eliminate CFCs.46

As a class, HCFCs suffered from acute regulatory uncertainty for several years, in the United States and internationally. Although HCFCs were the first alternatives available and destroyed much less ozone than CFCs, they still destroyed a little—about 2 to 10 percent as much as CFCs. Their intermediate level of harm made the right treatment of them unclear: Should uses be switched to HCFCs as fast as possible, or wait for entirely nonozone‐depleting options? Both producers and users worried that regulators would push them rapidly into HCFCs, then cut HCFCs as soon as CFCs were gone, not allowing long enough product lives for HCFCs or associated equipment to provide an adequate return on investment.47 And in fact, (p.182) calls to restrict HCFCs began to circulate in both the U.S. Congress and Protocol negotiations in early 1990. These proposals sought to restrict either the quantity or the duration of global HCFC use, some calling for product lifetimes as short as 10 years. By 1993 the regulatory treatment of HCFCs had been clarified as “transitional” chemicals, to be allowed product lifetimes sufficient to support both producer and user investments, but eliminated thereafter.

Despite all these risks, the new markets promised large rewards to the producers who picked the right chemicals, developed them successfully, built capacity in time to capture market share, and avoided early regulatory restrictions. These producers would enjoy strong positions in new markets with stronger barriers to entry and higher prices and profitability than for CFCs. Although the physical volume of the new markets would be smaller, prices of HCFCs were projected to be two to three times higher than pre‐Protocol CFCs, and patented HFC blends as much as 10 times higher.

Managing these risks, in particular the interactions between the expected success of new chemicals, their likely environmental impacts, and regulations, became essential elements of corporate strategy. Producers naturally sought to develop chemicals that were likely to escape strict regulation, lobbied to secure favorable regulation of those they decided to pursue, and attempted to exploit regulatory loopholes.48 All producers, for example, resisted 1990 calls for early HCFC phase‐outs, but DuPont—which had the strongest commitment to the chemicals—took more forceful measures. DuPont suspended construction of four HCFC plants worldwide in June 1990, and only resumed work two months later, when satisfied that Protocol negotiators and U.S. regulators would not cut HCFC lifetimes too short.49 After spending three years defending long HCFC product lifetimes, ICI reversed course in 1991, citing regulatory uncertainty in the EC, and announced—as did Hoechst the same year—that they would concentrate on HFCs and not attempt to commercialize any new HCFCs.50

Producers also sought to influence evolving regulations to cultivate their environmental images and to shape markets to their advantage. For example, once ICI had committed to a predominantly HFC strategy, it joined environmentalists and European regulators in advocating strict controls on HCFCs. Equipment manufacturers with little or no exposure to HCFCs also joined in advocating their early phaseout.51 Similarly, as DuPont reduced its CFC production, it twice advanced its CFC phaseout date—to year‐end 1996 in an October 1991 announcement, then to 1994 in a 1993 announcement—thereby keeping its own schedule ahead of that enacted in successive Protocol amendments.52 These repeated advances angered equipment manufacturers and other users, who claimed that DuPont was polishing its environmental image and yielding to unwarranted pressure from environmentalists, while they were bearing the costs in premature equipment conversion.53

In fact, the costs and risks borne by many users in the transition to new chemicals were significantly greater than those borne by CFC producers. For producers, the anticipated higher prices and greater concentration of alternative markets represented a benefit that might offset the risks of the transition, but for users these factors only made matters worse. Moreover, no viable alternatives were apparent for some sectors: the only plausible fluorocarbon alternatives for solvent uses of CFC‐113 all had serious toxicity problems or were judged too ozone‐depleting.54 Users also bore much of the risk of market disruptions, and of uncertain availability, cost, and performance of alternatives. These risks were dramatized in mid‐1992, when producers' announcements of rapid CFC production cuts before most alternatives (p.183) were in commercial production provoked widespread hoarding, price spikes, shortages, and plant shutdowns.55 Although shortages eased later in 1992, users still faced the unattractive choice between remaining dependent on high‐cost CFCs of uncertain availability, or moving rapidly to immature alternatives of uncertain performance, availability, cost, and regulatory lifetime. The next section discusses users' responses to these risks, which came to represent the largest risk to producers as well: the large‐scale loss of CFC markets to nonfluorocarbon alternatives.56

7.3 Mobilization of Chlorofluorocarbon User Industries: The Engine of Innovation

Prior to the 1987 signing of the Protocol, industries using CFCs had played little role in national or international policy debates. Although the Alliance represented U.S. producers and users and gained political influence from this breadth, its agenda was primarily shaped by a few large firms, the CFC producers and the manufacturers of refrigeration equipment. Between 1980 and 1987, discussion of CFC alternatives concentrated almost exclusively on chemical alternatives, with the strong presumption that desirable alternatives should resemble the CFCs as closely as possible in all relevant respects, and would come from current suppliers. The extremely low price elasticities estimated for CFCs as late as 1986 reflected the widespread assumption that users had few options to reduce their consumption.57 Although regulators began suggesting in 1986 that these assumptions and the weak participation of users in policy debates posed the risk of neglecting other alternatives, the only user group to participate in pre‐Protocol debates was the American Electronics Association. This group opposed the Protocol, arguing that there was no alternative for the CFC‐113 solvents on which its members depended.58

In late 1986, however, a few major users began seeking ways to reduce their consumption and emissions of CFCs through in‐house research programs and task forces.59 The Alliance provided early support to efforts, beginning with a November 1986 conference at which 500 producers and users began to investigate in detail where CFCs were used, where they were emitted, and how emissions could be reduced.60 Although the focus was on reducing emissions, these inquiries naturally revealed many ways to reduce use as well.

Within a year, as the desperate situation of many CFC users became clear, these preliminary investigations grew into a flood of development work by both current users and third parties seeking to market new solutions. The restrictions already agreed upon in the Protocol and the risk of further restrictions on CFCs and alternative chemicals gave users an interest in reducing their dependence on all these chemicals as quickly as possible.61 Further motivation to avoid fluorocarbons arose from the uncertainty and cost associated with the alternatives being developed by the CFC producers, and the perception that the CFC suppliers, in endorsing the movement to new chemicals that would harm users but might benefit them, had betrayed their customers. While some sectors were already reducing CFC use for unrelated reasons, these factors combined to spur an intense search for CFC alternatives, particularly nonfluorocarbon alternatives.62 By early 1989 most users were not just trying to meet the Protocol's 50 percent target, but to reduce their dependence on CFCs as rapidly as possible, while many were also trying to eliminate HCFCs.63 An early landmark in the rush to substitutes was the first CFC and Halons Alternatives Conference, jointly sponsored by the EPA, Environment Canada, (p.184) and the Conservation Foundation in January 1988, at which several critical breakthroughs were announced. Repeated annually with UNEP and the Alliance joining as cosponsors, the Alternatives Conference became the focal point for exchange of technical information about both fluorocarbon and nonfluorocarbon alternatives for each specific use.

These efforts made CFC elimination much faster and easier than projected, and made markets for HCFCs and HFCs substantially smaller than projected, although the pace and character of developments differed substantially among uses. Many important alternatives had been known or used at small scale for decades, while others required significant innovations. In some uses attractive non‐CFC alternatives appeared so fast that CFCs disappeared in a few years, while eliminating CFCs in others proved much more difficult than initially expected.

Aerosols, the largest remaining global CFC use in 1987, were eliminated the most easily and predictably. Despite a decade of claims to the contrary, the hydrocarbon propellants and alternative packaging formats that a few countries had adopted in the 1970s were immediately available in the rest of the world. Where hydrocarbon propellants were restricted because of inflammability or smog formation, dimethyl ether and HFC‐152a were also immediately available, although at substantially higher cost.64 Metered‐dose medical inhalants, a specialized class of aerosols, appeared much harder to substitute, but represented only a few percent of total aerosol use. The national aerosol bans of the 1970s had exempted these as essential uses, a precedent that the Protocol followed until substitutes were developed and tested in the late 1990s. As had been widely predicted, eliminating CFCs in aerosols—except for the last few percent—was easy, fast, and cheap.

Plastic foams used 267 Kte of CFCs in 1986, about one‐quarter of world use.65 Foams were a diverse group of uses, comprising three large types (flexible foams, rigid polystyrene, and rigid polyurethane) and several smaller types, which used different blowing agents and differed markedly in their ease of reducing CFCs. Some eliminated all fluorocarbons within three years, while others experienced persistent difficulties for more than 10 years. Flexible foams, used for cushioning in automobiles, mattresses, and furniture, were blown principally with carbon dioxide but used either CFC‐11 or methylene chloride as a secondary blowing agent to make the foam soft. For these foams, two approaches to reducing CFC use were evident by 1989: process changes to replace CFC‐11 with methylene chloride, and eliminating the secondary blowing agent entirely and accepting a slightly firmer foam. Union Carbide announced a new non‐CFC process in September 1988, while British foam manufacturers agreed in May 1989 to switch entirely to methylene chloride by 1993. European manufacturers reached a similar agreement with the EU Commission in August 1990. Although flexible foams were initially expected to be a large market for HCFCs 123 and 141b, the U.S. and European markets both essentially disappeared by 1994 as all users switched to nonfluorocarbon blowing agents.66

Rigid polystyrene, used in several forms to package food and other products, was mainly blown with CFC‐12, but some products also used hydrocarbons. Several forms of polystyrene switched quickly to HCFC blowing agents.67 One form, rigid polystyrene sheet, initially faced a more uncertain conversion because it needed regulatory approval for use in fast‐food packaging and food‐store display trays. Although it was only 3 percent of CFC use, this product's high visibility made it a target for consumer pressure as early as 1987.68 After McDonald's, the largest user, responded to early pressure by announcing in August 1987 that it would require (p.185) all its packaging material to be CFC‐free within 18 months, several packaging manufacturers worked rapidly to change their products. One firm had already been testing new blowing agents, and within two weeks announced a new process that could be implemented rapidly with little change of equipment, using a blend of HCFC‐22 and pentane. At the urging of EPA officials, the FDA gave expedited review to the new process and approved it for fast‐food packaging—although not for food‐store packaging—in October 1987. In January 1988, DuPont began marketing a highly purified HCFC‐22 for food packaging, and the largest U.S. maker of disposable food packaging announced it would switch all its operations to the new process within 60 days.69 EPA officials then convened a negotiation among packaging manufacturers and environmental groups in February, to promote rapid industrywide adoption of the new processes.

After the FDA's March approval for food‐store packaging removed the largest remaining obstacle, the food‐service packaging industry (FSPI) announced an industrywide voluntary commitment to end CFC use by the end of 1988. Because the new processes raised costs significantly, no firm could make the move alone but the whole industry could readily do it together. Three major environmental groups endorsed the plan, on the condition that the industry would switch to fully nonozone‐depleting blowing agents when available.70 The new technologies were shared domestically through the FSPI, and internationally through the Foams Technical Options Committee of the Protocol's Technology Assessment Panel, although their approval in food packaging was delayed in several countries due to lingering concerns over potential carcinogenicity. The promised second switch was completed within a few years, when manufacturers converted to hydrocarbons, HFC‐152a, and carbon dioxide.71 The polystyrene foam sector was anomalous in that it easily changed technology twice within a few years, since the cost of each change was so low.

The largest foam sector, and the largest single use of CFC‐11 in countries that had banned aerosols, was rigid polyurethane foam, used as insulation in buildings, appliances, and refrigerated transport. Eliminating CFCs here was most difficult, because CFCs represented a large share of total product cost and because alternative blowing agents were substantially less effective insulators. Consequently, switching to non‐CFC blowing agents would either raise the foam's cost or degrade its insulating value so much that the industry risked losing market to other insulating materials. Through 1988 there were serious concerns that this industry might not survive the wait for HCFCs to be commercialized; their difficulties were so acute that they were the only usage sector granted a partial exemption from the tax imposed on CFCs in the United States in 1990. By late 1989, however, it began to appear feasible to reduce CFC use in two stages: first by modifying blowing processes, then by switching to HCFCs, principally HCFC‐141b, as they became available in the mid‐1990s. When HCFC‐141b became available faster than projected, it largely replaced CFCs by 1993, although the industry struggled for several years to maintain insulation effectiveness by modifying blowing technology to obtain smaller cells. The challenges of this sector persisted, as it later also faced the greatest difficulties in moving away from HCFCs.72

The refrigeration and air‐conditioning sector, like the foam sector, comprised several distinct uses with diverse needs. This sector faced several unique difficulties in reducing CFCs: the amount and variety of long‐lived, CFC‐dependent capital equipment in service; the need for refrigerants to be compatible with machinery, lubricants, seals, hoses, and gaskets; and the need to meet regulated energy‐efficiency (p.186) standards. Long equipment lifetimes posed particularly hard choices, between making an immediate long‐term commitment to HCFCs that might not be available for servicing through the equipment's life, and staying with CFCs—which were highly likely to become unavailable—while awaiting nonozone depleting solutions.73

The main categories of refrigeration uses were automotive air‐conditioning, household refrigeration, commercial refrigeration, and the chillers used to air‐condition large buildings. The clear early leader to replace CFC‐12 in automobile air‐conditioning was HFC‐134a, although the need for equipment redesign and new lubricants and seals made it uncertain how soon a large‐scale switch could be accomplished.74 The major automakers regarded the conversion problems as serious enough that they pursued four other backup options before committing to HFC‐134a in 1991.75 They produced the first cars with HFC‐134a systems in 1992 and converted all models by 1994. In contrast to the early view that HFC‐134a systems would carry a 5–15 percent energy penalty, optimization of the new systems that took advantage for the first time of the thermodynamic properties of the lubricants gave efficiencies better than those of conventional systems by 1992.76 A diverse set of nonfluorocarbon systems was proposed that used hydrocarbons, helium, carbon dioxide, metal hydrides, and other refrigerants, but none of these received serious consideration. The reasons advanced for dismissing these alternatives were all legitimate concerns—for instance, inflammability, energy efficiency, retooling needs, and high electrical loads—but provided no clear basis for rejecting them out of hand in comparison with HFCs.77

The problem of eliminating CFCs in auto air‐conditioning was not limited to new cars, but included servicing the stock of 140 million vehicles on the road with CFC‐12 systems. Most automotive refrigerant was used not to charge new systems but to refill existing systems after leaks, service, or accidents, and it appeared that the supply of recycled CFC‐12 could meet less than half this need. Since HFC‐134a could not be used in older systems without hundreds of dollars of system refitting, many “drop‐in” replacements were quickly marketed to fill this gap.78 These included a few HCFC and HFC blends that matched the performance of CFC‐12 very closely, and many products of high inflammability or suspect performance.79 Although the EPA eventually named only two acceptable alternatives for automobile air‐conditioning—HFC‐134a and one blend—delay in issuing the rule allowed many potentially dangerous and damaging substitutes to be marketed, provoking enough concern that the Society of Automotive Engineers petitioned the EPA to restrict the sale of alternatives. A few years of technical refinements eased the problem of converting existing systems to HFC‐134a, with retrofit costs falling from about $1,000 to $100. Trends in coolant prices also helped ease the shift, as CFC‐12 rose from $10 to $50 per kilogram between 1992 and 1996 while HFC‐134a dropped from $25 to less than $10.80 Still, concern about shortage of CFC‐12 to service the existing automobile fleet was strong enough that in December 1993 the EPA asked DuPont to continue producing CFC‐12 through the end of 1995, although the company had planned to end production one year earlier.81

Household refrigerators were a small CFC use but, like foam food packaging, a highly visible one. Conventional home refrigerators used CFC‐12 as a refrigerant and were insulated with polyurethane foam panels blown with CFC‐11. The major U.S. manufacturers introduced new HFC‐134a models in 1993 and 1994.82 Although the change of refrigerant and the loss of CFC‐blown insulation initially posed large challenges to meeting energy efficiency targets required in 1993, design (p.187) changes allowed the targets to be met and exceeded.83 In fact, a competition sponsored by the EPA and 24 electric utilities stimulated development of competitive new CFC‐free models using 25–50 percent less energy than even the new standards required. The competition's $30 million prize went to Whirlpool for a model using HFC‐134a refrigerant and HCFC‐141b in the insulating foam.84 European household refrigeration went a completely different route, led by the efforts of the environmental group Greenpeace. It was widely known that light hydrocarbons such as propane and butane were feasible alternatives to CFC refrigerants, but had two disadvantages: pure hydrocarbons would require equipment redesign in most applications, and they were inflammable. Hydrocarbons were not seriously considered in the United States mainly due to concern about fire risk, although this may have been exaggerated in view of the small quantities used. But a Greenpeace campaigner in Germany learned of an independent inventor's hydrocarbon design, and in 1992 funded a failing East German appliance manufacturer, DKK‐Scharfenstein, to produce a household refrigerator based on the design. The refrigerator used a mixture of propane and isobutane as a refrigerant, and foam insulation blown with cyclopentane (in thicker‐than‐normal walls, to compensate for the lower thermal insulation value of this foam). Greenpeace and Scharfenstein introduced the refrigerator, which they called the Greenfreeze, at a July 1992 press conference and solicited direct orders from consumers. Although the initial model was a small refrigerator with no freezer, its promoters received 65,000 orders within four weeks. This immediate success averted the planned closure of the firm and led to its purchase by an international group of investors in November 1992.85 In March 1993, Green‐freeze won the German Blue Angel award for environmentally superior products.86

Although the major appliance manufacturers had declined Greenpeace's initial proposal and subsequently attacked the Greenfreeze as unsafe, consumer and public pressure soon forced them all to introduce their own hydrocarbon models.87 Bosch‐Siemens and Liebherr introduced hydrocarbon models in 1993 to compete with Greenfreeze. By the end of 1993, several firms had made design improvements that allowed larger hydrocarbon refrigerators with freezers, which were more energy‐efficient than CFC‐based models. After Greenpeace attacked them for releasing only a new HFC‐134a model in February 1993, Electrolux introduced a hydrocarbon model, and announced in August 1994 that it would convert its entire European line to hydrocarbons by the end of 1995.88 The other major manufacturers on the European continent, although not in the United Kingdom, followed with similar targets to switch most or all of their lines.89 Rapid penetration of household refrigerator markets by hydrocarbons followed in much of the industrialized world, and in major developing‐country markets.90 By the late 1990s, more than 12 million hydrocarbon refrigerators had been sold worldwide, and more than 100 hydrocarbon models were offered. The only market not using them was North America, where larger refrigerators posed harder technical problems and U.S. litigiousness made manufacturers more cautious about even minimal fire risk.

Commercial refrigeration systems in food stores and warehouses used three refrigerants, CFC‐12, HCFC‐22, and R‐502 (a blend of CFC‐114 and HCFC‐22), depending on the required operating temperature. The sector first expanded use of HCFC‐22, but a replacement for R‐502 in the coldest applications remained unavailable until 1993—a situation that led the Bush administration in its last days to reconsider its earlier request for rapid voluntary CFC production cuts.91 As in many other sectors, however, substitution eventually went faster than expected, following the introduction of several near drop‐in blends in 1993. At first there (p.188) were only HCFC blends for many applications, but HFC blends were commercialized for most applications by 1995 and HCFC use in the sector began to decline thereafter.92

Nonfluorocarbon refrigerants, such as hydrocarbons and ammonia, are also feasible in commercial systems, although the large charges of these systems and the need for charging in the field make fire and toxicity more serious concerns than for small household systems. Environmental groups promoted these options vigorously, but industry interest only began to grow only in the mid‐1990s with the introduction of systems using a tightly confined ammonia loop and a secondary loop of brine to carry cooling long distances.93 In 1994 and 1995, prompted in part by Greenpeace attacks for using HCFCs and HFCs, two U.K. supermarket chains announced plans in 1995 for stores using ammonia or hydrocarbon systems for all refrigeration.94 Secondary‐loop systems using ammonia also saw revived interest in shipping by 1994.95

The final major refrigeration use was CFC‐11 in chillers, which chill water to provide cooling in commercial buildings (home air‐conditioners run on HCFC‐22). After initially investigating multiple alternatives, some chiller manufacturers switched their lines to HCFC‐123, others to HFC‐134a. Both choices posed moderate technical problems—HCFC‐123 was not compatible with standard seals and gaskets, while HFC‐134a would not mix with standard mineral oils—but chiller manufacturers solved these, and by 1992 were offering designs that could operate with either conventional or new refrigerants.96 Some of these new designs represented major advances, offering improved performance, increased energy efficiency, and unprecedented low leakage levels.97 However, broader building design changes that would have integrated lighting improvements to reduce cooling loads, or more ambitious schemes to integrate heating, cooling and cogeneration applications with absorption‐cycle cooling cycles, did not make significant early inroads.98

In sum, nearly all refrigeration and air‐conditioning applications went to HCFCs and HFCs, including several blends. After an early proliferation of untested options caused confusion and delays in some sectors, the industry association issued codes, a numbering system, and standards for new refrigerants in mid‐1992.99 Mobile air‐conditioning and home refrigerators in the U.S. switched to HFC‐134a; commercial refrigeration went to HCFC‐22 and several HCFC and HFC blends; and chillers went to HCFCs 22 and 123, and HFC‐134a, depending on their operating pressure. Nonfluorocarbon alternatives did not receive serious consideration—except in household refrigerators, where Greenpeace's aggressive promotion of a hydrocarbon design eventually transformed the market everywhere except in North America. Nonfluorocarbon options began to gain renewed attention in the mid‐1990s, however, as potential alternatives for some uses that appeared to be the most dependent on HCFCs.

The solvent sector provides the most striking instances of users innovating around fluorocarbons, turning what was initially viewed as the hardest sector from which to eliminate CFCs into the easiest. CFC‐113 is an effective solvent, but is mild enough to use on plastic and rubber parts or for cleaning fine leather and suede, and sufficiently nontoxic to use for hand cleaning of precision metal parts. Methyl chloroform was a good substitute for many applications, and was widely used for metal degreasing, but also contributed to ozone depletion (although only about 10 percent as much as CFCs). Replacement of other solvents such as perchloroethylene and trichloroethylene, which the EPA listed as hazardous air pollutants (p.189) in 1986, had further increased growth of both CFC‐113 and methyl chloroform.100

Early attempts to find nonozone‐depleting alternatives to fluorocarbon solvents had limited success. HCFC‐132b initially appeared promising, but was too toxic. DuPont announced a few new solvents in 1988 that reduced CFC‐113 use by blending it with other chemicals, but the reductions were small (ranging from a few percent to 37 percent for various applications).101 In March 1989 both DuPont and Allied introduced several new blends based on HCFCs 141b and 123, but none of these adequately matched the performance of CFC‐113 for all applications. Most were stronger solvents, and so could not replace CFC‐113 in important electronics applications.102 Through 1987 and 1988, electronics and chemical firms continued to state that the prospects for replacing CFC‐113 solvents were poor, and that the EPA had overestimated them in enacting regulations.103

The first major change in this picture was announced at the January 1988 Alternatives Conference. AT&T had been working to reduce use of all chlorinated solvents since the late 1970s, for reasons of environment, cost, and workplace health, and had replaced them with aqueous cleaning of chips and boards at several plants by 1984. Aqueous cleaning was known to be suitable for nearly all nonsurface‐mount applications, but was not suitable for the more demanding surface‐mount applications. At the 1988 Conference, AT&T and a small firm, Petroferm, jointly announced a new, environmentally benign solvent, based on terpenes, that cleaned surface‐mount applications better than CFC‐113 for about the same price.104 The two firms had jointly developed processes for large‐scale electronics cleaning after an AT&T engineer saw the solvent advertised in a petrochemical‐industry supply magazine.105 AT&T announced that the new solvent would let them cut their use of CFC‐113 by one‐third, or about 450 tonnes per year.106 Petroferm also developed terpene solvents for a wide range of other plastic‐ and metal‐cleaning applications.107

The Solvents Technical Options Committee (TOC) under the Protocol was a major engine of innovation in reducing CFC solvent uses. Stephen Andersen, who was responsible for industry cooperative programs in the EPA's Stratospheric Protection Branch and was later named chair of the Solvents TOC, began approaching major solvent‐using firms at the January 1988 Alternatives Conference. He contacted both AT&T and Nortel, competing manufacturers of telecommunications equipment, to request their participation in the Solvents TOC and challenge them to reduce CFCs in their own operations. Both firms were already concerned about tighter CFC restrictions and were trying to reduce their use rapidly. Nortel committed publicly in July 1988 to cutting their worldwide CFC use in half by year‐end 1991, and later revealed that their actual goal was to eliminate them completely by that time. AT&T announced in 1989 that it would eliminate CFC emissions by the end of 1994.108

The Solvents TOC brought together major solvent‐using firms for its work during the first half of 1989. With top technical experts on specific solvent uses collaborating, the TOC found that in addition to identifying existing reduction opportunities, they were able to solve technical problems that had obstructed implementation of other reduction opportunities and even identify new opportunities. In one striking example, a group of TOC experts on a March 1989 site visit observed an innovative German machine that soldered inside a controlled‐atmosphere chamber. Although the machine's operation was problem‐ridden, the (p.190) group recognized that the concept of soldering in an inert atmosphere could control oxidation and reduce the need for fluxes, thereby making it unnecessary to clean components after soldering and allowing huge reductions in solvent use. Participating experts from AT&T, Nortel, and Ford immediately arranged to have their firms purchase the machines, and began collaborating to perfect the concept.109

From their experience on the Solvents TOC, officials from AT&T and Nortel became convinced that pooling effort and sharing technical information could greatly advance efforts to reduce CFC use. Working with the EPA's Andersen, they identified counterparts from other firms to invite to an organizational meeting in October 1989, immediately before the second Alternatives Conference, where they discussed forming an industry cooperative to continue such cooperation. Fourteen firms attended and nine agreed—despite suspicion about the EPA's intentions—to establish the cooperative, named the Industry Cooperative for Ozone Layer Protection (ICOLP).110 The EPA had already supported the formation of similar research consortia for halon reduction, for lubricant and equipment compatibility of automotive air‐conditioning alternatives, and for toxicity and environmental testing of new fluorocarbons, and so was able to guide the formation of ICOLP under the National Cooperative Research Act.111

By 1992, ICOLP had expanded to 15 corporate members and had developed affiliate memberships for government agencies and associations, including the U.S. Air Force—whose support was crucial when the group realized that more than half of world CFC solvents were used because U.S. military specifications required them.112 ICOLP's approach was based on R&D collaboration and free sharing of knowledge and technology with both members and outsiders. In addition, member firms produced technical manuals and an on‐line technology database to publicize their experience, and ran technology‐transfer workshops and provided consulting services to help eliminate ozone‐depleting solvents in developing countries. Members reported gaining several benefits from participating: consultation with top experts from multiple firms to help solve their hardest reduction problems; access to specific technologies, including some from their strongest competitors; pooling the risk of being unable to eliminate certain uses; and favorable publicity and access to government policy makers.113

Viable solvent options proliferated rapidly, spurred by the research initiatives of large, sophisticated firms like Nortel and AT&T, the review of promising technologies conducted by the Solvents TOC, and the collaborative research and free exchange of results promoted by ICOLP. By 1990, many diverse options (including aqueous, semi‐aqueous, hydrocarbon, and no‐clean) were available to replace virtually all CFC‐113 use. Although methyl chloroform was initially considered the best alternative to CFC‐113, many of the new nonfluorocarbon alternatives were also effective replacements for it.114 The corporate leaders who had pledged to eliminate CFCs from their operations all met their goals between 1991 and 1993, spurred by a May 1993 EPA deadline for warning labels on products made with ozone‐depleting substances. By late 1993, more than 60 major firms based in five countries had ended CFC use in their worldwide operations.115

This rapid innovation by solvent users took CFC producers by surprise. The one fluorocarbon that appeared promising as a CFC‐113 substitute, HCFC‐225, faced a long development time during which user innovation rapidly eroded its market.116 ICI, one of only two firms that marketed both CFC solvents and solvent cleaning equipment, recognized the inevitable in 1990 and began undercutting its own fluorocarbon business by selling terpene cleaning equipment.117 The solvents sector, (p.191) which in 1987 was thought the hardest place to eliminate CFCs, turned out to be among the easiest, and the fastest to move away from fluorocarbons altogether.118

The final usage sector restricted by the Protocol was the halons, which the 1987 Protocol froze at present levels beginning in 1992. Halons were highly effective fire extinguishers, but were substantially more ozone‐depleting than the CFCs.119 Although ozone depletion from bromine‐containing chemicals had been periodically discussed since 1975, halons were added to the Protocol's agenda only in the final months. Learning of their involvement in ozone depletion very late, the firms that made halons but not CFCs first tried, without success, to raise the same scientific questions about bromine—such as possible tropospheric sinks, uncertain stratospheric chemistry, and the need for years of measurements to verify atmospheric lifetimes—as had been discussed for chlorine for 15 years.120

Halon users took a different approach. The fire‐protection industry association (the National Fire Protection Association, or NFPA) learned that halons were ozone depleters only in 1986, when the EPA contacted them as they were about to revise the fire code to require full‐discharge testing of all halon systems. Gary Taylor, chair of the NFPA Standards Committee, pressed successfully to stop the proposed change, and the NFPA joined the EPA, the U.S. Air Force, the fire‐prevention equipment manufacturers, and three trade associations in establishing a group to identify and reduce unnecessary halon emissions. The EPA and the U.S. Air Force established a parallel government interagency group. These new networks, both established before the Protocol, provided technical support for the effort to include halons in the Protocol.121

These groups initially sought to limit halon emissions as much as possible to fighting fires. They quickly learned that although only 20 percent of annual halon production was released (as opposed to 85 percent of annual CFC production), less than 5 percent of releases went to fight fires. The great majority of emissions came from discharge‐testing of building fire systems. Even where fire codes did not require discharge testing, insurance companies often did, both to test system performance and to ensure that rooms were sealed tightly enough to hold the required halon concentrations. By early 1988 the two groups began developing and promoting alternative testing methods that did not require releasing halons, such as discharge‐testing with alternative chemicals or using blower‐doors to create a pressure differential.

The U.S. military was both a large halon user, consuming 35 percent of the U.S. market, and a strong leader in reducing it.122 In early 1989, the U.S. Air Force issued a new policy that limited halons to critical uses (i.e., in combat or other situations where evacuation was impossible) where no alternative provided adequate fire protection. The U.S. Marines, Navy, and Air Force then collaborated to develop the first halon recycling equipment, which was subsequently commercialized by private firms. The U.S. Army, which needed halons to protect armored‐vehicle crews in combat fires, adopted a strategy that combined replacing halons in noncombat uses and halon recycling, for which they led a government‐industry effort to develop a purity standard. The U.S. military also led international reductions, sponsoring two NATO conferences to share knowledge with other armed forces and to seek their support for accelerated phaseouts.123

These efforts rapidly identified several ways to make large reductions in halon emissions. First, large reductions were available simply by avoiding unnecessary releases, through such measures as alternative system testing methods. Second, sharing information among users revealed that a great deal of halon was used unnecessarily (p.192) due to mistaken beliefs that other extinguishing agents would damage valuable equipment. For example, a December 1989 workshop revealed that both IBM and NASA used water and CO2 systems to protect their computers without damage, while many other users believed that computers could be protected only with halons.124 Third, it was recognized that the large halon “bank,” 200–350 Kte sitting in tanks in existing systems, could be jointly managed to supply critical systems for many years without producing any more.125 Finally, several chemical halon substitutes were proposed beginning in 1990, although none fully matched halons' performance. Great Lakes Chemical announced a substitute in May 1990 that could serve most applications, but was too toxic for total flooding of occupied rooms and had an ODP of 0.5.126 In June 1990, DuPont proposed HCFC‐123 and HFC‐125 as somewhat costlier and less effective replacements for halons 1211 and 1301, respectively, pending toxicity testing and EPA approval under the SNAP program.127

It was because halons were believed extremely difficult to reduce that they were only frozen, not reduced, in the 1987 Protocol. These innovations in reducing unnecessary discharge and better managing of existing stocks reduced the need for halons so rapidly, however, that they were the first group of ozone‐depleting chemicals to be eliminated when production ceased at the end of 1993. Halons' unique usage patterns, the large existing stock, and the small fraction of consumption actually used where it mattered, allowed production to be eliminated long before chemical alternatives were fully commercialized. In that sense, the phaseout of halons was an even more extreme success story than that of solvents.

Although this user‐driven innovation occurred worldwide, it was most vigorous in North America—as was the movement away from CFCs in general. In Japan, the major initiatives were developed by industry associations with government guidance. In coordinated 1989 announcements, the automakers stated that they would switch to HFC‐134a, the cosmetics firms that they would switch to hydrocarbon propellants, and the electronics firms that they would phase out CFC‐113, although the means to do so were not yet clear.128 European nations remained well ahead of their Protocol deadlines, principally due to rapid aerosol cuts, but European firms remained far behind their North American counterparts in the more challenging sectors, with Germany a partial exception.129 While the major U.S. electronics firms had aggressive CFC‐elimination programs in place by 1989, surveys of U.K. electronics firms in 1990 and 1991 found that few had near‐term reduction goals and 20 percent planned to keep using CFCs as long as possible.130 Speaking at the 1992 Alternatives Conference, the German Environment Minister criticized German appliance firms for resisting his calls for a 1993 phaseout and announced a series of government‐industry consultations to determine how fast a phaseout was technically feasible.131 In December 1992, the U.K. Department of Trade and Industry was so concerned that firms were not adequately preparing for the impending EC cuts—an 85 percent CFC cut in January 1994 and a phaseout one year later—that it launched an industry awareness program to alert them.132 European industry lagged so far behind in 1993 and 1994 that an ICI spokesman criticized European refrigeration manufacturers for their lack of progress.133 The U.K. dry‐cleaning sector made so little advance preparation to abandon CFC‐113 that firms found themselves in a last‐minute panic on the eve of the phaseout in 1994.134

(p.193) 7.4 Regime Formation and Industry Strategy

In addition to being a time of rapid regime formation, the period from late 1986 to early 1988 marked a sharp divide in industry's response to the ozone issue. After 10 years of unchanged policy positions and no significant progress in identifying ways to reduce use of ozone‐depleting chemicals—despite sustained environmental concern—this period began a burst of development and innovation from producers and users of ozone‐depleting chemicals that made their early elimination appear plausible by 1989, and confidently foreseen by 1991. This shift to rapid progress was not spurred by any particular technological breakthrough. There was no technical reason that the burst of innovation that began in 1986–1988 could not have happened several years earlier. Indeed, many of the alternatives that allowed rapid reductions were partly known 10 or more years earlier, but needed sustained attention to refine them and solve associated problems before they could be viable. That this did not happen earlier represented a serious lost opportunity. That it happened at this time is an indication of the power of regulation and institutions to influence the rate and character of technological change.

Before negotiation of the Protocol, there was no market or regulatory pressure to promote the search for ways to reduce CFC use. The major CFC producers had investigated chemical alternatives in the 1970s as a defense against the risk of regulation, but abandoned the efforts when the immediate risk declined and they concluded that the new chemicals could not compete with CFCs in price. Their strategy was defensive, seeking to learn enough to identify alternatives in case of regulation, but to avoid revealing—or even learning—so much about alternatives as to make regulation more likely. Users of CFCs, regarding stringent restrictions as unlikely and trusting the producers to develop alternatives if cuts were imposed, had no interest in pursuing CFC reductions unless the same measures served more immediate competitive purposes, such as improving product quality or reducing costs. In aggregate, regulators never seriously considered controls, in part because they believed they would be too costly, while the firms best able to develop alternatives made no effort to do so because they believed regulation was unlikely. Their lack of effort in turn ensured that controls continued to appear too costly—a perception to which they were happy to contribute.

The linked conditions sustaining this no‐progress equilibrium were shattered by the Protocol's control measures—indeed, by the prior widespread shift in late 1986 to the belief that controls were likely. The first shift in industry's policy position in the fall of 1986 was pushed by rhetorical factors. The justifications on which industry had founded its opposition to controls were undermined by revelations about the 1970s alternatives research programs, resumed CFC market growth, and the 1986 assessment's conclusion that continued growth would likely bring large ozone loss. These factors compelled a change in industry groups' positions if they were to maintain their standing as responsible participants in the policy debate. But once this change was made, it contributed—along with other actors' responses to the same information—to a transformation of the strategic landscape for CFC producers and users. For the first time, it appeared likely that serious restrictions on CFCs might be enacted.

This shift fundamentally reshaped the interests of firms dependent on CFCs. The producers immediately revived their alternatives research programs, to defend against the risk of CFC restrictions and to pursue opportunities in alternatives markets once CFCs were restricted. The shift in users' interests was even stronger, (p.194) although differing among sectors. Users faced acute risks from the CFC cuts already agreed upon, and from the prospect of more stringent and broader controls. All users faced threats of disruption to their operations, degradation of product quality, and increased costs: for some, the survival of their business was threatened. With serious disruption unavoidable, users had an interest in reducing their dependence on CFCs as rapidly as possible, and the opportunity and incentive to reexamine their products and processes comprehensively to achieve this reduction. It was by no means clear that the best way to achieve this reduction was to adopt the chemical alternatives being promoted by the CFC producers. The unavoidable disruption of normal operations made it feasible to search more broadly for alternatives, and users had little reason for continued confidence that the CFC producers would protect their interests. The chemical alternatives were clearly going to cost more than CFCs, their performance and availability were uncertain, and they also faced risks of future restrictions. Consequently, many users tried to reduce their dependence not just on CFCs, but on all ozone‐depleting chemicals. The interests of producers and users, which had been aligned as long as the status quo could be reliably defended, now diverged—because the risks to users associated with alternative chemicals compelled them to look out for themselves, and because the prospect of consolidation in alternatives markets introduced directly opposed interests between producers and users on the issue of price.

The rapidity of innovations to reduce CFCs, the central role of user‐driven innovation, and the widespread development of nonfluorocarbon alternatives were all facilitated by domestic and international policy choices. The organization of the Protocol's Technology Panel encouraged innovation in all usage sectors, and its exclusion of CFC manufacturers differentially directed this advantage toward nonfluorocarbon alternatives. United States regulators were instrumental in alerting user firms to the coming restrictions, challenging them to take leading roles in reducing their usage, and facilitating the formation of collaborative organizations for firms to work together on shared technical problems. These bodies allowed joint technical work among competitors without the risk of antitrust charges, and also allowed groups of competitors to take decisions in parallel that carried cost penalties, not imposing a competitive disadvantage on any one for moving first. Crucially, officials also worked with industry to identify and remove barriers to the development of alternatives that arose from government regulations and practices.

The results of these efforts differed strongly among usage sectors, in the ease and rapidity of the initial movement from CFCs, the share moving to other halocarbons—and later, in the ease of those that initially went to halocarbons in moving beyond them. In aggregate, reducing CFC use was much faster and easier than had been projected. By late 1990, it was evident that nearly every sector had numerous technical opportunities available to reduce use, and that new options were appearing monthly. By 1992, it was clear that CFC use could be reduced virtually to zero in all sectors without serious disruption.135 The solutions that enabled this were diverse, including conservation measures, new fluorocarbons, nonfluorocarbon chemicals, and process changes that made the services formerly performed by CFCs unnecessary. Alternatives of all these kinds were available earlier and performed better than projected, but the sectors that appeared most difficult shifted over time: first solvents and halons, then insulating foams, and finally certain refrigeration equipment.

The degree of cooperation among firms also differed strongly among usage sectors. Solvents and halons, the sectors that first appeared most difficult, had two (p.195) important advantages: firms had substantial capability to solve technical reduction problems if they cooperated; and because these are standards‐dominated industries, firms were willing to cooperate. Because firms needed to meet well‐defined performance standards but did not compete to outperform the standard, it was relatively unlikely that a firm could gain an important competitive advantage by finding an alternative to CFCs and not sharing it. In renouncing CFC alternatives as a dimension of competition, firms gave up only modest commercial opportunities tangential to their main lines of business. Similar circumstances facilitated cooperation in food‐packaging foams. The industry was standards‐dominated because of the need for FDA approval, and feasible alternatives were quickly and obviously available, at a cost penalty that was not so large as to make the entire industry uncompetitive with other products. Consequently, firms could readily agree to make the move together, with minimal competitive effects. In contrast, there was little cooperation among firms in the refrigeration and insulating foams sectors. In both these sectors, alternative chemicals posed large, important uncertainties about product performance, which could confer decisive competitive advantage on a firm that made a significant breakthrough. Moreover, for insulating foams, not even the whole industry could move together to adopt an alternative with a cost or performance penalty, because the penalties appeared large enough to risk loss of the entire market to other insulating materials.

The intensity of users' rush from CFCs and the flood of nonfluorocarbon alternatives developed by users and third parties greatly exceeded producers' expectations. The largest chemical producers expected to profit from the switch to chemical alternatives that they would control. In DuPont's successive estimates of the shares

                   Industry Strategy and Technical Innovation, 1987–1992

Figure 7.2 Projected and Realized Replacements For CFC Uses. It Was Initially Projected That About 40 Percent Of the CFCs Used In Industrialized Countries Would Be Replaced By HCFCs Or HFCs. As the Phaseout Was Implemented, However, Attractive New Nonfluorocarbon Alternatives To CFCs Were Repeatedly Identified. Ultimately, HCFCs and HFCs Replaced Only About 20 Percent Of Former CFC Use. Source: UNEP 1999b.

(p.196) of former CFC markets going to various alternatives, the share predicted to go to other fluorocarbons stood at 39 percent in mid‐1989, rose slightly to 41 percent in 1991, then plummeted to 26 percent in 1993. In the actual replacement of CFC uses realized by 1998, only 20 percent went to HCFCs and HFCs.136

Producers responded to these changes in various ways. While some exited the business, others shifted to smaller volumes of higher‐margin chemicals or broadened their business to sell related products and services in addition to chemicals, such as equipment, technical advice, and recovery and recycling facilities.137 Those that remained in the fluorocarbons business needed several elements in their post‐Protocol strategy. They of course tried to pick the most profitable alternatives to develop, to produce them cheaply, to market them effectively, and to defend a reputation for good corporate citizenship. But in addition, once they had invested in producing particular chemicals, they had to defend these as long as possible against both regulatory restrictions and competition from other alternatives. In part this required attempting to elicit regulatory commitments before making investments. Since large‐scale technology choices were being made both in markets and in regulatory arenas, it also sometimes required attacking the safety or environmental acceptability of other proposed alternatives, as well as their performance—especially those that appeared the most threatening.138 Producers who chose not to invest in particular alternatives had clear interests in joining environmentalists or governments in highlighting their environmental harms and calling for their stringent regulation, particularly when marketing competing products.

The political strategy of industry associations such as the Alliance also had to change. The Alliance's strategy of uniting producers and users around shared interests in the original chemicals became infeasible as soon as the chemicals came under serious threat. This approach remained valid only for a narrow set of issues on which the interests of producers and users remained sufficiently aligned, such as defending adequate product lifetimes for fluorocarbon alternatives and, more delicately, defending continued availability of CFCs for long enough to let all usage sectors make an orderly transition to new technologies. Both these stances implied resisting the most aggressive environmentalists and regulators, whose interest in reducing ozone depletion as rapidly as possible made them willing to risk premature abandonment of early investments in interim technologies as less ozone‐depleting ones become available. But having retreated from their long‐standing opposition to any restrictions, the Alliance and other industry bodies faced the subtler challenge of continuing to defend their members' interests while maintaining their standing as responsible participants in continuing regulatory debate.


(1.) Worldwide CFC revenues were about $3 billion. U.S. production was about 380 Kte, more than 1 kg per capita, generating about $700 million of revenue (Cogan 1988).

(2.) Hammitt et al. 1986, p. 11 figs. 1.1, 1.2.

(3.) European CFC exports were 127 Kte in 1985 (ENDS 152 [9/87]: 23).

(4.) UNEP 1989f.

(5.) NYT 9/17/87, p. A12; Chemical Week 9/30/87.

(6.) Glas 1989, p. 144.

(7.) The most recent analysis had found reduction opportunities of only 20 to 40 percent in the United States (Camm et al. 1986). Even in September 1987, when a display of alternative technologies was assembled in Montreal to show delegates the promising state of reduction opportunities, the technologies on display were highly limited and experimental (Andersen, Morehouse, and Miller, 1994; Andersen, Buxton, and Taylor interviews).

(8.) Manzer 1990, p. 31; Chemical Week 142 no. 14 (4/6/88): 7.

(9.) J. Steed, C&EN 11/24/86.

(10.) Congressional testimony of E. Blanchard, DuPont, 5/13/87.

(11.) Chemical Week 9/30/87, p. 6, quotes estimates of J. Glas, DuPont.

(12.) International Chlorofluorocarbon Substitutes Committee 1986, p. 2–2; Lagow interview.

(13.) DuPont press release and letter to L. Thomas, 1/5/88 (88‐2‐9). Four other producers joined over the following year.

(14.) IER 2/10/88, p. 110; Chemical Engineering 1/18/88, p. 22.

(15.) Von Schweinichen 1989; UNEP 1989f, p. 127.

(16.) AFEAS 1995.

(17.) In Mar. 1989, a senior DuPont researcher said that banning HCFC‐22 would be “a catastrophe. Everyone is counting on it as a substitute.”NYT 3/7/89, p. C1.

(18.) ENDS 152 (9/87): 152.

(19.) ENDS 157 (2/88): 4.

(20.) ENDS 156 (1/88): 9–11.

(21.) ENDS 157 (2/88): 5; ENDS 160 (5/88): 3–4; ENDS 171 (4/89): 29; European industry comments on the cost of the aerosol ban in 82‐4‐3, 83‐1‐26, 85‐2‐1, 86‐4‐1.

(22.) Environmentalists were divided over this substitution, because it made a large immediate improvement but HCFC‐22's ODP of 5 percent was near the high end of proposed alternatives. A coalition of U.S. environmental groups supported the switch on the condition that a nonozone‐depleting blowing agent be adopted as soon as possible. But since Friends of the Earth UK switched its consumer campaign to foam food packaging—including that blown with HCFC‐22—after its success on aerosols, FOE was briefly caught in the awkward position (p.329) of having its U.S. organization support this use while its British organization was boycotting it (ENDS 158 [3/88]: 15–17; ENDS 163 [8/88]: 7; New Scientist 5/26/88, p. 56).

(23.) Those with no hydrogen have long atmospheric lifetimes; those with too much hydrogen are inflammable; and those with intermediate shares of H and Cl but not enough F tend to be toxic (McLinden and Didion 1987).

(24.) DuPont and ICI quotes in NYT 3/31/88, p. D1; Chemical Engineering 1/18/88, p. 22; Glas 1989; DuPont, Fluorocarbon/Ozone Update 7/88 (88‐2‐5).

(25.) Two others that initially looked promising (HCFCs 132b and 133a) were found in early 1987 to be too toxic (M. Jones, “In Search of the Safe CFCs,”New Scientist 5/26/88, p. 56), and three more (HCFCs 22 and 142b, HFC‐152a), were poorer matches for 11 because of their low boiling points but were thought to be useful in blends (Dishart, Creazzo, and Ascough 1987).

(26.) DuPont, Fluorocarbon/Ozone Update 7/88 and 12/88 (88‐2‐5); SRI International 1995.

(27.) UNEP 1989e.

(28.) ICI switched a pilot plant from HCFC‐141b to 123 in April 1990, while DuPont canceled a 141b production plant in 1993 (Chemical Week 12/19/90, p. 51; Bradley 1994).

(29.) Draft regulations published March 1993, final regulations December 1993 (58 FR 65018, 12/10/93).

(30.) Fluorocarbon/Ozone Update 7/98, 12/98; IER 1/30/91; Cox and Lesclaux 1989; Zellner 1989. (Note, however, that Atochem announced expansion of HCFC‐141b in France and the United States as late as 1991.)

(31.) Two others (HCFC‐142b and HFC‐152a) had suitable properties and were already produced at small scale for polymer production, but were inflammable. Other candidates (HCFC‐124, HFCs 125 and 143a) had less suitable properties but could be used in blends (DuPont, Fluorocarbon/Ozone Update 3/87 [87‐2‐1]).

(32.) NYT 3/31/88, p. D1; Daily Telegraph (London) 12/23/88, p. 8.

(33.) Business Week 5/17/93, p. 78.

(34.) These plans doubtless included some tactical overstatement to deter others from entering. Chemicals and Materials 5/13/91; Financial Times 3/13/91, p. 4.

(35.) DuPont press release 7/92 (92‐2‐9); CMR 11/23/92, p. 5; Chemical Week 4/17/91, p. 8.

(36.) CMR 12/23/92, p. 5; ENDS 246 (7/95): 15–16.

(37.) UNEP 1991b, pp. 4‐1–4‐3.

(38.) IER 7/15/92, p. 471; ENDS 209 (7/92): 6.

(39.) Chemical Week 12/18/91, p. 27; Appliance 49, no. 2, (2/92): 42; Manzer 1990; Didion and Bivens 1990.

(40.) Allied Signal press release 5/28/96 (96‐2‐2); C&EN 71, no. 11 (3/15/93): 5–6; IER 5/ 4/94, p. 390.

(41.) Chemical Week 4/26/89, p. 25.

(42.) Kaiser and Racon made no attempt to develop substitutes, and announced plans to exit the market immediately after the Protocol. Pennwalt tried to market HCFCs 141b and 142b, both of which it already produced in small volumes for specialized applications (Chemical Engineering 1/18/88, p. 22; Chemical Week 4/26/89, p. 25).

(43.) SRI International 1995, p. 543.7001 I.

(44.) For example, after Trane redesigned its large chillers to operate on HCFC‐123 and Carrier switched to HCFC‐22 and HFC‐134a, the firms attacked each other's choice of chemical so persistently that DuPont's vice chairman publicly called on them to stop, stating that their attacks were encouraging users to delay choosing either and to stay with CFCs (Krol 1992, reported in Appliance 49, no. 2 [2/92]: 42).

(45.) Having accepted the need to eliminate CFCs, producers repeatedly had to resist proposals to do it faster than they believed possible. For example, after the 1989 assessments reported it would be feasible to eliminate CFCs in 10 years, bills were immediately introduced to eliminate them in five (Chemical Week 3/1/89, p. 16).


(46.) Chemical Week 12/18/91, p. 27.

(47.) CMR 8/31/92, p. 9, quotes a refrigerant distributor: “There is no question in my mind that government bodies will aim their guns at HCFCs within a year after the CFC phaseout is complete.”

(48.) E.g., Benedick 1998, p. 103; Faucheux and Noel 1988.

(49.) F. A. Vogelsburg, “If society decides it doesn't want these chemicals, we're not going to waste our time producing them” (quoted in Financial Times, 6/22/90, p. 22); ENDS 184 (6/90): 4; IER 7/90, pp. 313–14; CMR 6/25/90, p. 7; NYT 6/23/90, p. 31; IER 11/7/90, p. 459; Chemical Week 12/19/90, p. 51.

(50.) ICI Chemicals and Polymers, “The Ozone Issue and Regulation: An ICI Appraisal,” June 1990 (90‐2‐3).

(51.) ENDS 198 (7/91): 4; ENDS 209 (6/92): 6, noted that ICI produced only one HCFC, and would benefit if competitors with stronger commitments to the chemicals were subjected to tight controls. See also BNA, International Environment Daily, 3/1/93; GECR 3, no. 9, (5/ 3/91).

(52.) IER 11/6/91, p. 591; WSJ 3/9/93, p. B5.

(53.) CMR 3/15/93, p. 1.

(54.) Three HCFCs (141b, 132b, and 225) were considered, and many firms switched to HCFC‐141b, especially in Europe, before it was restricted due to its high ODP (CMR 3/15/ 93, p. 1; Andersen interview).

(55.) CFCs 11 and 12 were available only at a $1 to $2 premium over their posted prices of about $3/lb (WSJ 4/29/92, p. B1). The panic was exacerbated by President George H. W. Bush's Feb. 1992 announcement of an advance of U.S. phaseouts, in which he asked producers to cut output by half immediately.

(56.) ENDS 158 (3/88): 15, quotes an ICI official that high alternative costs will prompt users to look elsewhere, to “less effective products which may not even be made by the chemical industry”; CMR 7/9/90 discusses reliability problems and equipment failures associated with early use of HCFC‐123 in chillers.

(57.) D. Pearce, “The EC Approach to Control of CFCs” (86‐4‐1) estimated price elasticity for CFCs 11 and 12 as 0.2; Camm et al. 1986 estimated 0.01 to 0.03 for CFC‐11, 0.01 to 0.07 for CFC‐12, and 0.1 to 0.16 for CFC‐113.

(58.) American Electronics Association, “The Electronics Industry, CFCs, and Stratospheric Ozone Reduction,” submitted to hearings of House Subcommittee on Natural Resources, Agriculture, Research, and Environment, 3/10/87 (pp. 303–17); Rome workshop report, Mar. 1986 (86‐1‐2); Canadian delegation report (86‐1‐1).

(59.) Firms with early internal programs included DEC, IBM, GM, Ford, and McDonalds (WSJ 12/2/86, p. 4).

(60.) Testimony of R. Barnett, Senate Committee on Environment and Public Works, 1/28/ 87; WSJ 12/2/86, p. 4.

(61.) The threat of further cuts to CFCs and other fluorocarbons was present by Jan. 1988. HCFCs were at risk for their contribution to ozone loss; HFCs, for their contribution to global climate change (e.g., Doniger 1988b).

(62.) For example, CFC use in auto air‐conditioning was sharply reduced in the 1980s by design changes that cut the average charge from 4 to 2.5 pounds and average losses from 25 to 10 percent per year (IER 4/8/87, p. 164).

(63.) Moore 1990.

(64.) Drug and Cosmetic Industry 148, no. 6 (6/91): 18.

(65.) UNEP 1989c, p. 1

(66.) ENDS 165 (10/88): 4; ENDS 172 (5/89): 7; ENDS 187 (8/90): 24; Bradley 1994.

(67.) WSJ 5/13/88, p. A23.

(68.) Cook 1996b, p. 21.

(69.) IER 2/10/88, p. 109.

(70.) “Industry Announces Voluntary Phaseout Program,” Foodservice and Packaging Institute press release, 4/12/88 (88‐2‐10); NRDC, EDF, FOE, “Statement of Support for the (p.331) FSPI's Fully Halogenated CFC Voluntary Phaseout Program,” 4/12/88 (88‐2‐6); DuPont, Fluorocarbon/Ozone Update 7/88, p. 7 (88‐2‐5).

(71.) Cook 1996b, p. 26.

(72.) Hammitt et al. 1986, p. 31; ENDS 175 (8/89): 7; Plastics World 11/89; Journal of Commerce 6/25/90; Plastics World 3/91, pp. 93–95; CMR 1/18/93, p. 3; Plastics World (1/93): 44–50; UNEP 1994c.

(73.) Supermarket Business 6/92; Buildings 11/92, p. 70; Bradley 1994.

(74.) For example, a 1987 consultant's study projected that none of the Protocol's first reduction in 1994, and only 2 percent of the second cut in 1999, would come from this sector (Atkinson 1989: conversion problems were so severe that large‐scale implementation of HFC‐ 134a by the mid‐1990s “would be a major achievement”).

(75.) They considered HCFC‐22, which was used in the 1950s but whose high vapor pressure required heavy hoses and fittings; HCFC‐142b, which Pennwalt promoted vigorously but the automakers viewed with skepticism because of its inflammability; and two HCFC blends (Andersen 1989).

(76.) All models produced in North America, Europe, and Japan were converted by 1994, while those manufactured in developing countries continued to use CFC‐12 until after 2000 (UNEP 1994c; Andersen interview).

(77.) NYT 3/31/88, p. D1; Cogan 1988, pp. 111–12; Air Conditioning, Heating, and Refrigeration News 11/2/87; Appliance 49, no. 2 (2/92): 42; ENDS 233 (6/94): 26.

(78.) Chicago Tribune 5/23/93, p. 1.

(79.) Appliance 49, no. 2, (2/92): 42.

(80.) Seidel 1996.

(81.) NYT 12/19/93, p. 30.

(82.) Business Week 9/20/93, p. 85; NYT 2/21/1994, p. C6.

(83.) Gants 1988.

(84.) C&EN 7/5/93, p. 15; Cook and Kimes 1996, pp. 55–65.

(85.) IER 12/16/92, p. 835.

(86.) IER 3/24/93, p. 210.

(87.) IER 8/12/92, p. 521; The Guardian 11/19/92, p. 2.

(88.) IER 2/24/93, p. 123; ENDS 235 (8/94): 25.

(89.) ENDS 229 (2/94): 25.

(90.) UNEP 1994b, pp. 95–110; Ozone Action 19 (7/96): 7 reports China receiving grants from the multilateral fund for 31 hydrocarbon refrigeration projects.

(91.) CMR 1/18/93, p. 3.

(92.) Global Environmental Change Report (94‐5‐1).

(93.) UNEP 1994b, pp. 94–107; UNEP 1995b, p. EX. 4.

(94.) ENDS 240(1/95): 29–30; ENDS 247 (8/95): 29–30.

(95.) ENDS 232 (5/94): 29.

(96.) Appliance 49, no. 2, (2/92): 42; Smithart 1993.

(97.) E.g., a new line of Trane HCFC‐123 chillers announced at the 1992 Alternatives Conference reduced annual coolant losses from 25 to 0.5 percent (“Trane Announces New Ozone‐Friendly Chiller,”PR Newswire 10/8/92).

(98.) Andersen 1989.

(99.) Building Design and Construction 11/92.

(100.) Chemical Week 1/8/86, p. 56.

(101.) DuPont, Fluorocarbon/Ozone Update 7/88, p. 7; 12/88, p. 4 (88‐2‐5).

(102.) Chemical Week 4/26/89; R. N. Miller 1989; Allied Signal 1989.

(103.) American Electronics Association, White Paper submitted to House Subcommittee on Natural Resources, 3/10/87, pp. 303–17; NYT 3/31/88, p. D1; Daily Telegraph (London) 11/23/88, p. 8.

(104.) NYT 1/14/88, p. 15.

(105.) J. K. Johnson 1994.

(106.) IER 2/10/88, p. 109.


(107.) Hayes 1989.

(108.) Vice President Margaret Kerr revealed Nortel's true goal under prodding by Tolba at UNEP's Oct. 1988 meeting in The Hague (Wexler 1996, p. 93; Rose and Fitzgerald 1992; Boyhan 1992).

(109.) Wexler 1996 pp. 90–91.

(110.) In addition to Nortel and AT&T, founding members were Boeing, Digital Equipment, Ford, GE, Honeywell, Motorola, and Texas Instruments, all of which followed Nortel and AT&T's lead in announcing corporate phaseout or near‐phaseout goals (ICOLP, “A New Spirit of Industry and Government Cooperation,” 95‐2‐2).

(111.) NYT 10/11/89, p. A27; Boyhan 1992; Wexler 1996, p. 93.

(112.) The EPA and Department of Defense had formed an interagency working group in Mar. 1988, to test and evaluate alternative cleaning procedures as a first step to revising the specifications (Wexler 1996, p. 92; Keane 1995).

(113.) AT&T reported that eliminating 80 percent of its CFC use was easy, but that it really needed ICOLP's help for the last 5–10 percent; British Aerospace reported that one technology it gained from its archival Boeing, a lubricant used in riveting aircraft wings, was worth much more than their cost of participating (Machine Design 4/23/93, p. 60; GECR 4, no. 5, (2/26/93); ICOLP (95‐2‐2).

(114.) E.g., 3M and Petroferm announced a drop‐in terpene replacement for methyl chloroform in vapor degreasing at the 1992 Alternatives Conference. R&D 34, no. 14, (12/92): 28.

(115.) Ozone Action 8 (9/93): 4.

(116.) ENDS 184 (5/90): 16.

(117.) ENDS 182 (3/90): 8–9.

(118.) By 1993, an ICI official described the former CFC‐113 market as “totally fragmenting, with many of the alternatives not‐in‐kind” (i.e., nonfluorocarbon). International Environment Daily 3/1/93.

(119.) The Protocol defined the ODP of halon‐1301 as 3 and that of halon‐1211 as 10, relative to a value of 1 for CFCs 11 and 12.

(120.) B. Tullos, Great Lakes Chemical, testimony at EPA regulatory hearings, Jan. 1988.

(121.) The Halon Alternatives Research Corporation (HARC), like PAFT, AFEAS, and ICOLP, received antitrust protection under the 1984 National Cooperative Research Act (Taylor, Morehouse, and Andersen interviews).

(122.) IER 1/15/91, p. 16.

(123.) Andersen, Morehouse, & Miller, 1994.

(124.) Mauzerall interview, quoting IBM official: “We don't use vacuum tubes in our computers anymore.”

(125.) Presentation of T. Morehouse, 1992 Alternatives Conference, summarized in R&D 34, no. 14, (12/92): 28.

(126.) The new chemical, CHF2Br, was marketed as Firemaster 100 (Mauzerall 1990, p. 30).

(127.) C&EN 73, no. 5, (1/30/95): 25; Business Insurance, 5/18/92, p. 23.

(128.) Nihon Keizai Shimbun, 9/2/89, in Schreurs 2001.

(129.) CSM 3/5/92, p. 6; IER 9/15/92.

(130.) ENDS 181 (2/90): 8.; IER 7/31/91, p. 422.

(131.) IER 2/26/92, p. 98.

(132.) Times (London) 12/29/92; IER 2/10/93, p. 110.

(133.) Chemical Week 2/17/93, p. 23; The Economist 1/29/94, p. 69.

(134.) ENDS 233 (6/94): 10.

(135.) Many participants in the 1992 Alternatives Conference said ending CFCs by 1995 was very likely feasible, a stark change from two years earlier, though not all uses had yet identified solutions (R&D 12/92, p. 28).

(136.) DuPont, Fluorocarbon/Ozone Update, 8/89 (88‐2‐5); CMR 7/9/90; F. A. Vogelsburg, speech, Mar. 12–13, 1991, London (91‐2‐4); Cook 1996a, p. 8, fig. 4; Bradley 1994; UNEP 1999b, p. 23.


(137.) ENDS 184 (5/90): 15.

(138.) For example, the intensity of attacks on Greenfreeze suggests concern with losing much more of the refrigeration sector to hydrocarbons (IER 2/24/93, p. 135).