Arkema Inc. v. Environmental Protection Agency

United States Court of Appeals FOR THE DISTRICT OF COLUMBIA CIRCUIT Argued May 10, 2010 Decided August 27, 2010 No. 09-1318 ARKEMA INC., PETITIONER v. ENVIRONMENTAL PROTECTION AGENCY, RESPONDENT Consolidated with 09-1335 On Petitions for Review of an Order of the Environmental Protection Agency Dan Himmelfarb argued the cause for petitioners. With him on the briefs were John S. Hahn, Roger W. Patrick, Brian D. Netter, William J. Hamel, David M. Williamson, and Gia V. Cribbs. David E. Mills was on the brief for amici curiae Professors Robert H. Gertner, et al., in support of petitioners and vacatur. 2 Perry M. Rosen, Attorney, U.S. Department of Justice, argued the cause for respondent. With him on the brief was Diane E. McConkey, Attorney. Before: SENTELLE, Chief Judge, BROWN, Circuit Judge, and RANDOLPH, Senior Circuit Judge. Opinion for the Court filed by Circuit Judge BROWN. Dissenting opinion filed by Senior Circuit Judge RANDOLPH. BROWN, Circuit Judge: In 1987, the United States signed the Montreal Protocol on Substances that Deplete the Ozone Layer (the Protocol). The agreement sought to “limit or eliminate the[] production and consumption of ozone depleting substances” by incrementally decreasing the manufacture or consumption of these substances using a series of decreasing caps, with an initial focus on chlorofluorocarbons (CFCs). Protection of Stratospheric Ozone; Listing of Substitutes for Ozone-Depleting Substances, 65 Fed. Reg. 42,653, 42,655 (proposed July 11, 2000). In 1990, the Protocol was amended to accelerate the phaseout schedule for CFCs and identified hydrochlorofluorocarbons (HCFCs) “as transitional substitutes for chlorofluorocarbons . . . and other more destructive ozone-depleting substances, but agreed to phase out HCFCs because of their significant potential to destroy stratospheric ozone as well.” Id. Under the terms of the Protocol, the United States was required to phase out 35% of its historic HCFC production (measured by 1989 levels) by 2004; 65% by 2010, 90% by 2015; 99.5% by 2020, and 100% by 2030. Protection of Stratospheric Ozone: Allowance System for Controlling 3 HCFC Production, Import and Export, 68 Fed. Reg. 2820, 2821 (Jan. 21, 2003) (2003 Rule). Section 607 of the Clean Air Act (CAA) requires the Environmental Protection Agency (EPA or the Agency) to use a market-based cap and trade regulatory system—a system of pollutant production allowances transferable between companies and between types of HCFCs—to control production and importation of HCFCs. See 42 U.S.C. § 7671f. In 2003, the EPA promulgated a final rule for a cap and trade regulatory system, allocating HCFC allowances on a one-time basis to each participating company and authorizing those companies to expend their baseline allowance during each control period (a calendar year). See 2003 Rule, 68 Fed. Reg. at 2823. The 2003 Rule allowed these companies to trade their allocations, subject to EPA approval, between companies and between regulated HCFCs on an annual or permanent basis. The EPA sought to create a system with “maximum flexibility,” id. at 2833, making “allowances easily tradable with minimum regulatory interference and oversight, thereby encouraging companies to make business decisions[s] as they would in an unregulated industry.” Id. at 2824. Preparing for the intermediate reduction in HCFC production in 2010 (the “2010 stepdown”), the EPA initiated a new rulemaking in late 2008. See Protection of Stratospheric Ozone: Adjustments to the Allowance System for Controlling HCFC Production, Import, and Export, 73 Fed. Reg. 78,680 (proposed Dec. 23, 2008) (Proposed Rule). The EPA outlined five possible approaches in the Proposed Rule, one of which was to continue the existing cap and trade system and reduce the caps pro rata. Id. at 78,687. In the Final Rule, however, the EPA chose to honor only inter- company transfers of baseline allowances and to disallow permanent baseline changes resulting from inter-pollutant trades. Protection of Stratospheric Ozone: Adjustments to the 4 Allowance System for Controlling HCFC Production, Import, and Export, 74 Fed. Reg. 66,412, 66,421–22 (Dec. 15, 2009) (Final Rule). Arkema Inc. (“Arkema”), Solvay Flourides, LLC, and Solvay Solexis, Inc. (“Solvay”) (collectively Petitioners) filed this consolidated action arguing the Final Rule is arbitrary and capricious and has an impermissibly retroactive effect as to their HCFC baseline allowances. We agree the Final Rule unacceptably alters transactions the EPA approved under the 2003 Rule, and we therefore vacate the Final Rule in part and remand it to the EPA. I A. The Clean Air Act In 1990, Congress enacted Title VI of the Clean Air Act, 42 U.S.C. §§ 7671–7671q, implementing as domestic law the Protocol’s goal of protecting the stratospheric ozone layer. Title VI established a framework for gradually phasing out the production and consumption of ozone-depleting substances by annually reducing quantities of CFCs (class I substances) and HCFCs (class II substances) as measured against a baseline year. For HCFCs, Title VI defined the baseline year as a representative calendar year selected by the EPA Administrator. See id. § 7671(2). Congress gave the Administrator substantial discretion, permitting the EPA to accelerate the phaseout if a more stringent schedule was deemed necessary or practicable. See id. § 7671e. Otherwise, the Administrator was generally authorized to promulgate rules providing for the issuance of allowances controlling the production of class I and class II substances and governing the transfer of allowances. See id. § 7671c (class I substances); id. § 7671d (class II substances); id. § 7671f (transfer of allowances). 5 In Section 607 of Title VI, Congress gave the EPA a single, clear directive concerning transfers of allowances: “Such rules shall insure that the [transfer] transactions under the authority of this section will result in greater total reductions in the production in each year of class I and class II substances than would occur in that year in the absence of such transactions.” Id. § 7671f(a). Subsection (b) of Section 607 (“Interpollutant transfers”) permits “a production allowance for a substance for any year to be transferred for a production allowance for another substance for the same year on an ozone depletion weighted basis.” Id. § 7671f(b)(1). Subsection (c) (“Trades with other persons”) permits “2 or more persons to transfer production allowances (including inter-pollutant transfers which meet the requirements of subsections (a) and (b) of this section) if the transferor of such allowances will be subject, under such rules, to an enforceable and quantifiable reduction in annual production which . . . exceeds the reduction otherwise applicable to the transferor . . ., exceeds the production allowances transferred to the transferee, and would not have occurred in the absence of such transaction.” Id. § 7671f(c). B. The 2003 Rule On January 21, 2003, the EPA promulgated regulations to ensure compliance with the first stepdown milestone, reducing HCFC consumptions by 35% and freezing production, by January 1, 2004. 2003 Rule, 68 Fed. Reg. at 2821. The individual company baselines were calculated using the company’s individual highest ozone depletion potential (ODP)-weighted consumption among the years 1989, and 1994 through 1997. Id. at 2832. The EPA believed selecting a company’s year of highest activity over a range of years as its baseline created less of a disadvantage to the industry and the HCFC market as a whole than basing each 6 company’s baseline on a single year. See id. at 2831–32. The EPA allocated calendar-year allowances equal to a percentage of the baseline for specified control periods (defined as the period from January 1 to December 31). To carry out the 1993 phaseout schedule, the EPA issued calendar-year allowances of 100% of baseline for HCFC-22 and HCFC- 142b for each control period from 2003 through 2009. See 40 C.F.R. § 82.16(a) (2003). The EPA noted it was allocating HCFC allowances “on a one-time basis.” 2003 Rule, 68 Fed. Reg. at 2823. Thus, “allocations would remain the same from control period to control period (one calendar year to the next) until each chemical is phased out or until the percentage of baseline allowances is reduced to ensure compliance with the Protocol cap. Only through permanent transfers of allowances would a company’s baseline allocation be changed.” Id. The 2003 Rule allowed both inter-pollutant and inter- company transfers of allowances. Id. at 2833–34; 40 C.F.R. § 82.23(a), (b) (2003). The preamble to the Rule distinguished between (1) the “permanent transfer of baseline allowances,” which it described as “a lasting shift of some quantity of a company’s allowances to another company,” and (2) “the transfer of current-year allowances.” 2003 Rule, 68 Fed. Reg. at 2835. The EPA explained that with a permanent transfer of baseline allowances, “[i]n all relevant subsequent years, the transferor’s quantity of baseline allowances would be permanently reduced, while the transferee’s quantity of baseline allowances would be permanently increased.” Id. (emphasis added). Furthermore, “at the time of a reduction step or a phaseout of the substance, the current holder of baseline allowances that were received in a permanent transfer would be the person who would have them deducted.” Id. The EPA stated it “w[ould] allow permanent transfers of baseline allowances with those allowances 7 disappearing at the phaseout date for the specific HCFC, regardless of what inter-pollutant transfers had taken place.” Id.1 The regulations described procedures for making inter- pollutant and inter-company transfers. See 40 C.F.R. § 82.23(a), (b) (2003). To satisfy section 607’s reduction mandate, the regulations applied an offset to every HCFC trade by deducting 0.1% from the transferor’s allowance balance. See id. § 82.23(a)(i)(G) (2003). Moreover, the regulations stated: “A person receiving a permanent transfer of baseline production allowances or baseline consumption allowances (the transferee) for a specific class II controlled substance will be the person who has their baseline allowances adjusted in accordance with phaseout schedules in this section.” Id. § 82.23(d) (2003). C. The 2010 Rule In anticipation of the 2010 stepdown, on December 23, 2008, the EPA issued a proposed rule adjusting the allowance system for control of HCFCs for the control periods 2010– 2014. Proposed Rule, 73 Fed. Reg. 78,680. As the EPA explained in the Proposed Rule, as of December 23, 2008, it had not yet allocated any calendar-year allowances for HCFC- 142b or HCFC-22 to cover the 2010 control period and beyond. See id. at 78,686. Without a grant of calendar-year allowances for those HCFCs, the EPA’s then-current regulations prohibited their production and import after December 31, 2009. Id. Under the 2007 Montreal Protocol Adjustment, as of January 1, 2010, the EPA needed to reduce by 65% its aggregate HCFC baseline for production and 1 The Final Rule indicates 26.1% of baseline for HCFC-22 and .47% of baseline for HCFC-142b are available in 2014. 40 C.F.R. § 82.16(a) (2009). Thus, neither HCFC will be completely phased out during the 2010–2014 stepdown period. It is unclear when the complete phaseout for these two HCFCs will occur, but all HCFCs must be phased out by 2030. 8 consumption. The Proposed Rule stated the EPA was proposing to apportion company-specific baselines in amounts equivalent to the existing baselines published in the 2003 Rule, see 40 C.F.R. §§ 82.17 (2003) (apportionment of baseline production allowances), 82.19 (2003) (apportionment of baseline consumption allowances), “adjusted as necessary to reflect permanent transfers of baseline allowances.” Proposed Rule, 73 Fed. Reg. at 78,686. The EPA presented five options for allocating HCFC-22 and HCFC-142b allowances. The first was allocating a percentage of the baseline production and consumption allowances “with or without considering any permanent baseline transfers and/or inter-pollutant transfers that resulted in a different amount of production or consumption for a specific HCFC.” Id. at 78,687. The Proposed Rule also contained a table showing the EPA’s proposed apportionment of production and consumption baselines for HCFC-141b, HCFC-22, and HCFC-142b. Id. at 78,694 tbl. As the EPA acknowledged, the table “reflect[ed] adjustments resulting from approved inter-pollutant and/or inter-company transfers of baseline allowances (i.e., permanent rather than calendar- year allowances).” Id. at 78,693. EPA noted only transfers that had occurred prior to June 16, 2008 (second quarter of 2008 control period) would be reflected in the final apportionment of baselines in the final rule. See id. The proposed amendments to the regulations also updated the baseline production and consumption allowance tables to reflect permanent inter-pollutant transfers of baseline allowances. Id. at 78,703–04 (proposed amendments to 40 C.F.R. §§ 82.17, 82.19). On December 15, 2009, the EPA issued the Final Rule. Final Rule, 74 Fed. Reg. 66,412. The EPA stated it was updating the baselines for HCFC-22 and HCFC-142b to 9 reflect permanent inter-company baseline transfers but would not recognize permanent intra-company, inter-pollutant transfers of baseline allowances. Id. at 66,421. Noting only two companies had supported recognizing such transfers, the EPA stated that if it had recognized the transfers, those two companies “would receive 38% and 912% more HCFC-22 allowances while the remaining companies would each receive 16% fewer HCFC-22 allowances.” Id. The EPA concluded recognizing the transfers might “disrupt the entire market in 2010 and . . . encourage greater disruption in future control periods.” Id. “[A]djusting the baselines to reflect intra-company, inter-pollutant transfers,” the EPA explained, “could create incentives for future manipulation of the allocation system in anticipation of future control periods.” Id. The EPA further noted, “[C]onsidering the language of section 607 and the legislative history, EPA believes that section 607(b) is best read as permitting only year-by-year inter-pollutant transfers.” Id. Accordingly, the EPA stated it “interprets section 607 as requiring that all inter-pollutant transfers, whether occurring between companies or within a single company, be conducted on a yearly—and thus temporary—basis.” Id. at 66,422. The EPA represented it had made statements consistent with its new position and noted the percentages of baseline allowances for HCFC-22 and HCFC-142b in the table in 40 C.F.R. § 82.16 (“Phaseout Schedule of Class II Controlled Substances”) had changed from the Proposed Rule due to the Agency’s decision not to recognize intra-company, inter-pollutant baseline transfers. Id. at 66,422, 66,428–29. By not accounting for those transfers, the total allocation in the Final Rule decreased. Id. at 66,428. 10 II This court may review any final action of the EPA promulgating national air quality or emissions standards. 42 U.S.C. § 7607(b)(1). We will set aside a final rule promulgated under the CAA if it is “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law,” 42 U.S.C. § 7607(d)(9)(A), or “in excess of statutory jurisdiction, authority, or limitations, or short of statutory right,” id. § 7607(d)(9)(C), and must affirm the Agency’s final action if the record shows all relevant factors were considered and the agency articulated a “rational connection between the facts found and the choice made.” Catawba County, N.C. v. EPA, 571 F.3d 20, 41 (D.C. Cir. 2009). At its core, this is a dispute over whether the Agency has changed its interpretation of Title VI of the CAA. Petitioners argue the EPA’s practice under the 2003 Rule differs markedly from the policy outlined in the Final Rule and thus Petitioners accuse the EPA of departing from its prior policy without adequate explanation. The EPA responds by insisting it “never declared that inter-pollutant transfers . . . would be recognized in subsequent step-down regulations,” EPA Br. at 25, and even if it did change its policy, it “adequately explained the reasons for choosing not to recognize past inter- pollutant transfers in establishing baselines for the new regulatory period,” id. at 30. Of course, the Agency is entitled to change its mind as long as its new direction falls within the ambit of its authorizing statute and the policy shift is adequately explained. The requirement that an agency provide a reasoned explanation for its actions ordinarily means the agency must “display awareness that it is changing position.” FCC v. Fox Television Stations, Inc., 129 S. Ct. 1800, 1811 11 (2009). But while an agency must show good reasons for its new policy, there is no requirement that the policy change be justified by reasons more substantial than those the agency relied on to adopt the policy in the first place. See id. at 1810. Thus, it would seem a straightforward proposition for the EPA to state the Final Rule departed from the policy it had adopted in the 2003 Rule and explain its reasons for doing so. Instead the Agency attempts an awkward straddle. On the one hand, the EPA insists the Final Rule did not change an established policy. On the other, the Agency argues that even if the Final Rule did alter a long-standing policy, the Agency’s reasoned explanation cures any defect. The Agency asserts “Congress left it to the broad discretion of EPA to determine how transfers of baselines are to be treated.” EPA Br. at 48. This is true, and that fact entitles the Agency to Chevron deference, but it does not allow it to retroactively alter the consequences of its actions. As we have lamented, the retroactivity rules are easy to state, less easy to apply. Generally, an agency may not promulgate retroactive rules without express congressional authorization. See Bowen v. Georgetown Univ. Hosp., 488 U.S. 204, 208 (1988). The EPA does not argue section 607 authorizes it to promulgate retroactive rules, and no such congressional intent is apparent from the statutory language. See also Sierra Club v. Whitman, 285 F.3d 63, 68 (D.C. Cir. 2002) (“The relevant provisions of the Clean Air Act contain no language suggesting that Congress intended to give EPA the unusual ability to implement rules retroactively.”). A rule operates retroactively if it takes away or impairs vested rights. See Nat’l Mining Ass’n v. United States Dep’t of Interior, 177 F.3d 1, 8 (D.C. Cir. 1999) (National Mining I) (quoting Ass’n of Accredited Cosmetology Sch. v. Alexander, 12 979 F.2d 859, 864 (D.C. Cir. 1992)). The critical question is whether the interpretation established by the new rule “changes the legal landscape.” Id. (quoting Health Ins. Ass’n of Am., Inc. v. Shalala, 23 F.3d 412, 423–24 (D.C. Cir. 1994)). If a new rule is “substantively inconsistent” with a prior agency practice and attaches new legal consequences to events completed before its enactment, it operates retroactively. Nat’l Mining Ass’n v. Dep’t of Labor, 292 F.3d 849, 860 (D.C. Cir. 2002) (National Mining II); see also Mobile Relay Assocs. v. FCC, 457 F.3d 1, 11 (D.C. Cir. 2006) (explaining “[r]etroactive rules ‘alter[] the past legal consequences of past actions’” (quoting Bowen, 488 U.S. at 219 (Scalia, J., concurring))). Even where a rule merely narrows “a range of possible interpretations” to a single “precise interpretation,” it may change the legal landscape in a way that is impermissibly retroactive. National Mining I, 177 F.3d at 141. Did that happen here? We think it did. As a threshold matter, the EPA’s own transfer allowance form, Form 2014.03, apparently allowed applicants to request inter-pollutant baseline transfers. Item 2.2 on Form 2014.03 asks the applicant to identify the “Type of Allowances Transferred” and specifically states the applicant must “check only one.” The two options are “Current Year Allowances” and “Baseline Year Allowances.” Arkema and Solvay both used Form 2014.03 to seek the EPA’s approval of their inter- pollutant baseline transfers from HCFC-142b to HCFC-22. Arkema submitted the form on April 18, 2008 and Solvay submitted it on February 15 and March 4, 2008. Both Arkema and Solvay checked “Baseline Year Allowances” under Item 2.2 and indicated they were seeking to transfer HCFC-142b baseline allowances to HCFC-22. In a series of “Non-Objection Notices,” the EPA approved each of these transfers. The Non-Objection Notices indicated the “Allowance Type” for each of the allowance transfers from 13 HCFC-142b to HCFC-22 was “Baseline.” These transactions thus support the Petitioners’ assertion that the EPA approved changes to their baseline allowances as a result of inter- pollutant trades on the same basis as changes resulting from inter-company trades. The EPA attempts to minimize the significance of its approval of Petitioners’ inter-pollutant baseline trades by suggesting an agency’s policies are “evidenced by its express statements, not by divining the agency’s purported thought processes from check boxes on forms,” and “[n]othing on the form or accompanying letters describes these transfers as applying in perpetuity.” EPA Br. at 26. However, these are not the only instances when the Agency’s practices supported Petitioners’ position. In anticipation of the Proposed Rule, the EPA sent Arkema letters on August 14, 2008 and January 9, 2009 confirming the Agency had recognized the transfers of inter-pollutant baseline allowances. The EPA tries to dismiss the subsequent letters by suggesting they “do nothing more than show that Petitioners’ prior inter-pollutant transfers were reflected in the proposed baselines for the 2010-2014 period.” EPA Br. at 27. Thus, the EPA argues, it “originally proposed to reflect inter- pollutant trades in the baselines for 2010-2014, but it ultimately chose not to do so.” Id. This is not an accurate reading of the letters, which purported to represent the EPA’s assessment of the current status of Arkema’s allowances. For instance, the EPA’s August 14, 2008 letter stated, “The table below compares: (1) the baselines for your company that are currently published in 40 CFR Part 82, Subpart A, as apportioned in [the 2003 Rule]; and (2) the current baselines for your company, updated to reflect permanent trades of baseline allowances pursuant to 40 CFR 82.23.” Letter from Ross Brennan to Dawn Mattia at 3 (Aug. 14, 2008) (emphasis 14 added). The table reflected Arkema’s inter-pollutant baseline transfers from HCFC-142b to HCFC-22 as the “Current baseline (as of July 1, 2008).” Id. at 3 tbl. Nothing in the letter indicated this was the EPA’s “proposed” baseline for the 2010–2014 period. To the contrary, the letter’s use of “current” and “permanent” strongly suggests the EPA considered the inter-pollutant transfers equivalent to inter- company transfers. Moreover, in the Proposed Rule, the EPA noted the “leading option” it was considering for implementing the 2010 stepdown was an approach that could account for inter- pollutant transfers: “Allocating a percentage of the baseline allowances (§§ 82.17 and 82.19) for each HCFC respectively with or without considering any permanent baseline transfers and/or inter-pollutant transfers that resulted in a different amount of production or consumption for a specific HCFC.” Proposed Rule, 73 Fed. Reg. at 78,687. And the Proposed Rule tabulated the baseline allowances for the Petitioners to include the inter-pollutant transfers. See id. at 78,693 tbl. The record thus reflects that the EPA’s practice under the 2003 Rule was to allow Petitioners’ baseline transfers of inter-pollutant allowances. Not only did the Agency approve inter-pollutant transfers where the companies indicated they intended the transfers to change their baselines, but the EPA also provided tables reflecting such baseline transfers in its calculations and included an option in the Proposed Rule that would have continued such transfers. The Agency’s approval and acknowledgment of Petitioners’ actions distinguishes this case from situations where a company’s unilateral business expectations are thwarted by a change in the regulatory framework. 15 Despite this clear practice, the EPA refused in the Final Rule to recognize Petitioners’ inter-pollutant transfers in their baseline allowances. The EPA discussed at some length the concerns that led to the change in policy—the potential for manipulation; the way recognition of permanent inter- pollutant transfers would shift the phasedown from a “worst- first” approach to an “ODP-weighted” approach; and the possibility of market disruption or distortion. Final Rule, 74 Fed. Reg. at 66,420–22. These reasons may shield the Agency’s prospective application of the Final Rule from an arbitrary and capricious challenge. However, the Final Rule is a successive iteration in a long-running regulatory regime, and the effect of the EPA’s refusal to include Petitioners’ inter-pollutant transfers in their baseline allowances is to undo what the EPA had, in practice, approved under the 2003 Rule. See National Mining II, 292 F.3d at 860 (“If a new regulation is substantively inconsistent with a . . . prior agency practice, . . . it is retroactive . . . .”). Indeed, the EPA’s fundamental justification for refusing to recognize the Petitioners’ inter- pollutant transfers was that section 607 of the CAA precluded it from doing so: “After considering the language of section 607 and the legislative history, EPA believes that section 607(b) is best read as permitting only year-by-year inter- pollutant transfers . . . . Hence, EPA interprets section 607 as requiring that all inter-pollutant transfers, whether occurring between companies or within a single company, be conducted on a yearly—and thus temporary—basis.” Final Rule, 74 Fed. Reg. at 66,421–22. Because the EPA’s interpretation of section 607 in the Final Rule contradicts its past practice, narrowing the range of options and altering the legal landscape, the Agency’s refusal to account for the Petitioners’ baseline transfers of inter- pollutant allowances in the Final Rule is impermissibly retroactive. See National Mining I, 177 F.3d at 8; see also 16 Mobile Relay Assocs., 457 F.3d at 11. The Agency contends it may interpret what constraints, if any, Congress placed on the establishment of baselines for each stepdown period. That’s true. The Final Rule may more accurately track the statutory mandate and better reflect the Agency’s commitment to a “worst-first” HCFC reduction strategy, and after reaching this realization, the Agency is certainly entitled to change its mind and institute a program that forbids baseline inter-pollutant transfers in the future. But once the Agency has approved permanent changes to the baseline as a result of inter-pollutant transfers on the same basis as changes resulting from inter-company transfers, it cannot, without Congress’ express authorization, use its new statutory interpretation to undo these completed transactions. The EPA argues businesses could not reasonably expect baseline changes to last beyond the particular stepdown period. But obviously they could expect exactly that outcome. In the Final Rule, the EPA carried the inter-company baseline changes forward, and carrying inter-pollutant baseline changes forward was one of the options the Agency itself proposed. III Relying primarily on the EPA’s present interpretation of the 2003 Rule and its preamble, the dissent insists the EPA did not alter its position. But as explained above, even affording the EPA the deference it is due, the Petitioners have clearly demonstrated from the record that the EPA’s interpretation of section 607 did change between the 2003 Rule and the Final Rule. Moreover, if the EPA’s position was always as settled as the dissent suggests it was, the Agency’s “leading option” in the Proposed Rule, which proposed recognizing inter-pollutant baseline transfers, would make little sense. The dissent also points to the cover letters to 17 Solvay’s 2008 transfer requests as demonstrating its position is “demonstrably false.” However, in determining what the EPA’s practice was, we find the statements and actions of the Agency itself to be the most telling evidence. In any event, Solvay’s cover letters are hardly persuasive support for the EPA’s current interpretation of section 607—that an inter- pollutant transfer may only be made on an annual basis. As the dissent acknowledges, Solvay’s cover letters referred to the transaction as a “multi-year transfer for 2008 and 2009.” The record reflects the EPA’s practice under the 2003 Rule was to approve inter-pollutant transfers as permanent changes to baseline allowances, a practice the EPA impermissibly attempted to undo in the Final Rule by changing its interpretation of section 607. Following the EPA’s approach in this case, the dissent next argues that even if the Agency did change its position, the change did not create a retroactivity problem. The dissent’s criticism sails wide of the mark. The Final Rule is impermissibly retroactive not because it unsettled Petitioners’ expectations or imposed new liabilities on past conduct but quite simply because it attempted to undo the Petitioners’ inter-pollutant baseline transfers based on the EPA’s new interpretation of section 607. Although the 2010 stepdown gave the EPA occasion to adjust its distribution of allowances, it did not give the EPA an opportunity to revisit the baseline transactions it previously approved. IV The Final Rule appears to have been properly promulgated under the Administrative Procedure Act, and we see no reason to vacate the Rule on that ground. Prospectively, the EPA can limit inter-pollutant trades to a single year and can prohibit inter-pollutant baseline transfers. 18 But the Final Rule cannot have retroactive effect. We therefore grant the petitions for review in part, vacate the Final Rule insofar as it operates retroactively, and remand the case for prompt resolution consistent with this opinion. So ordered. RANDOLPH, Senior Circuit Judge, dissenting: I do not agree that EPA changed its position. And I do not agree that if EPA had changed its position, its new rule would violate the law against retroactive regulations. EPA never stated, not once, that a company’s inter-pollutant transfers would permanently and forever alter the company’s baselines for these pollutants. EPA’s 2003 regulations said nothing of the sort. In fact, the regulations indicated otherwise. The only “permanent” baseline transfers EPA recognized were transfers from one company to another. EPA’s preamble to its 2003 regulations defined a “permanent transfer” as “a lasting shift of some quantity of a company’s allowances to another company.” 68 Fed. Reg. at 2835 (emphasis added). This passage should have alerted all affected parties that there was no similar option for transfers within one company, a position reflected in the regulations themselves. Under 40 C.F.R. § 82.23(d) (2003), a “person receiving a permanent transfer of baseline production allowances . . . will be the person who has their baseline allowances adjusted in accordance with phaseout schedules in this section.” As the government argues, this provision contemplates inter-company transfers for the simple reason that “there would be no reason to identify which party receives the baseline adjustment if it was a one-party transaction.” EPA Br. at 55. The majority never directly confronts the preamble and § 82.23(d). Instead, we are offered two indirect responses. The first is that if EPA had never authorized “permanent” inter-pollutant transfers, EPA’s proposal to incorporate those previous changes into the 2010 baselines “would make little sense.” Maj. Op. at 16. In fact, the opposite is true. If EPA had authorized and approved transfers that permanently and forever altered the petitioners’ baselines, the proposal to allocate allowances “with or without” considering those transfers would be nonsensical. 2 The majority’s second response relies on EPA’s standard transfer form. But the completed form, together with an agency official’s approval of the a transfer, signified only that EPA recognized a transaction known as a “baseline inter-pollutant transfer.” Nowhere did the EPA form address the effective duration of the transaction. The only correspondence to use the word “permanent” in connection with these trades specifically noted that it “should not be interpreted to signal a particular course of action for development of HCFC allowances for 2010- 2014.” Petitioners nevertheless insist that EPA led them to believe that their inter-pollutant transfers would permanently adjust their baselines. As to petitioner Solvay, this claim is demonstrably false. Each of Solvay’s 2008 transfer requests reads: “[T]his transfer is for Baseline Year Allowances and therefore is being done on a permanent basis (i.e. multi-year transfer for 2008 and 2009).” In other words, Solvay believed that a baseline inter- pollutant transfer was “permanent” only in the sense of lasting beyond the current year and applying to the remainder of the stepdown period (2008 and 2009 at the time of the request). Clearly, whatever statements EPA made did not convey to Solvay that the transfers were “permanent” in the sense petitioners now urge. Thus the only evidence of how the regulated parties actually interpreted the disputed terminology shows Solvay interpreting “baseline inter-pollutant transfer” to mean what EPA now says it meant. Instead of relying on Solvay’s actual transfer requests, the majority locates in EPA’s transfer form an unwritten implication that if baseline inter-company transfers applied beyond the end of the regulatory period, so did baseline inter-pollutant transfers. No reasonable company would have viewed the form’s arrangement of check-boxes as implying any such thing. Still less could a company reasonably base its investment decisions 3 on such flimsy evidence. EPA’s failure to interpret its form in the way the majority interprets it hardly amounts to arbitrary action. The majority owes EPA deference for its interpretation of its form, but it gives the agency none. See Global Crossing Telecomms., Inc. v. FCC, 259 F.3d 740, 746 (D.C. Cir. 2001); Wyoming Outdoor Council v. U.S. Forest Serv., 165 F.3d 43, 53 (D.C. Cir. 1999). The interpretive choice reflected in the 2003 regulations and the one EPA expressly adopted in the latest regulations was utterly predictable. If regulated parties were permitted to make permanent shifts in allowances between pollutants, the transactions would give rise to potential manipulation. Arkema attempted to convert permanently its HCFC-142b baseline allowances into HCFC-22 baseline allowances just before its HCFC-142b account was reduced (or “stepped down”) by 99.6 percent. If Arkema had succeeded, not only would it have increased its total allowances but, going forward, it would have been able to convert its new HCFC-22 allowances back into HCFC-142b allowances on an annual basis. This maneuver would have allowed Arkema to continue producing HCFC-142b while avoiding the 99.6 percent stepdown. EPA, recognizing that it had never promised to make these types of transactions available, expressly made the transactions unavailable. For these reasons, I do not believe the 2010 regulations altered the policy embodied in the 2003 regulations. As a result, there is no retroactivity problem. Cf. Health Ins. Ass’n of Am., Inc. v. Shalala, 23 F.3d 412, 424 (D.C. Cir. 1994). But even if EPA changed its position, the 2010 regulations barring future production and consumption of hydrochlorofluorocarbons based on previous inter-pollutant transfers is not retroactive. A rule is not retroactive solely because it “upsets expectations based in prior law.” Landgraf v. 4 USI Film Prods., 511 U.S. 244, 269 (1994). Rather, a retroactive rule is one that “takes away or impairs vested rights acquired under existing laws, or creates a new obligation, imposes a new duty, or attaches a new disability in respect to transactions or considerations already past . . ..” Id. (quoting Soc’y for the Propagation of the Gospel v. Wheeler, 22 F. Cas. 756, 767 (C.C.N.H. 1814)); accord Ass’n of Accredited Cosmetology Sch. v. Alexander, 979 F.2d 859, 864 (D.C. Cir. 1992). The majority relies primarily on cases in which new regulations imposed liability on conduct that did not give rise to liability at the time it occurred. See, e.g., Nat’l Mining Ass’n v. U.S. Dept. of the Interior, 177 F.3d 1, 8 (D.C. Cir. 1999); Health Ins. Ass’n, 23 F.3d at 425. These cases all recognize the “unfairness of imposing new burdens on persons after the fact.” Landgraf, 511 U.S. at 270. But when no new liability is created, it is not enough that a party relied on existing regulations in the hope that the law would remain unchanged. See, e.g., Mobile Relay Assocs. v. FCC, 457 F.3d 1, 11 (D.C. Cir. 2006); DirecTV, Inc. v. FCC, 110 F.3d 816, 826 (D.C. Cir. 1996); Ass’n of Accredited Cosmetology Sch., 979 F.2d at 864. This is why a new regulation that merely “affects a regulated entity’s investment made in reliance on the regulatory status quo” will be upheld as long as it is reasonable. Mobile Relay Assocs., 457 F.3d at 11. EPA’s 2010 regulations impose no new liability or duty on petitioners. They do not invalidate the effectiveness of the inter- pollutant transfers for past years or impose liability on the companies for exercising the allowances gained through them. For the 2010 regulations to be considered retroactive, they must have taken away some vested right petitioners possessed. But these companies had no vested rights. The 2003 regulations established allowances only for the years 2003 to 2009. 40 5 C.F.R. § 82.16 (2003). The preamble to the regulations made clear that EPA would award allowances by notice-and-comment rulemaking for years after 2009 and that it was merely “likely” that EPA would do so by allotting a percentage of the baselines established by the 2003 regulation. 68 Fed. Reg. at 2823. Even if the inter-pollutant transfers had been recognized as carrying over to years after 2009, that baseline would not vest these companies with the right to produce or consume any particular quantity of hydrochlorofluorocarbons—EPA would still have to grant allowances as a percentage of that baseline. The 2010 regulations may have frustrated the petitioners’ expectations that they would be able to produce and consume certain quantities of hydrochlorofluorocarbons. But like the regulation upheld in Mobile Relay Associates, the effect of the regulation is purely prospective. “To conclude otherwise would hamstring not only [EPA] in its [hydrochlorofluorocarbon] management, but also any agency whose decision affects the financial expectations of regulated entities.” Mobile Relay Assocs., 457 F.3d at 11. The majority recognizes that EPA adequately explained its decision in the 2010 regulations not to recognize permanent inter-pollutant transfers. Maj. Op. at 15. Because those regulations do not impair vested rights, create new obligations, impose new duties, or attach new disabilities based on past transactions, the regulations are not impermissibly retroactive and should be upheld.