Cook County, Illinois v. Chad F. Wolf

In the United States Court of Appeals For the Seventh Circuit ____________________ No. 19-3169 COOK COUNTY, ILLINOIS, et al., Plaintiffs-Appellees, v. CHAD F. WOLF, Acting Secretary of Homeland Security, et al., Defendants-Appellants. ____________________ Appeal from the United States District Court for the Northern District of Illinois, Eastern Division. No. 19 C 6334 — Gary Feinerman, Judge. ____________________ ARGUED FEBRUARY 26, 2020 — DECIDED JUNE 10, 2020 ____________________ Before WOOD, Chief Judge, and ROVNER and BARRETT, Cir- cuit Judges. WOOD, Chief Judge. Like most people, immigrants to the United States would like greater prosperity for themselves and their families. Nonetheless, it can take time to achieve the American Dream, and the path is not always smooth. Recog- nizing this, Congress has chosen to make immigrants eligible for various public benefits; state and local governments have 2 No. 19-3169 done the same. Those benefits include subsidized health in- surance, supplemental nutrition benefits, and housing assis- tance. Historically, with limited exceptions, temporary receipt of these supplemental benefits did not jeopardize an immi- grant’s chances of one day adjusting his status to that of a le- gal permanent resident or a citizen. Recently, however, the Department of Homeland Security (DHS) issued a new rule designed to prevent immigrants whom the Executive Branch deems likely to receive public as- sistance in any amount, at any point in the future, from enter- ing the country or adjusting their immigration status. The Rule purports to implement the “public-charge” provision in the Immigration and Nationality Act, 8 U.S.C. § 1182(a)(4). States, cities, and nonprofit groups across the country have filed suits seeking to overturn the Rule. Cook County, Illinois, and the Illinois Coalition for Immi- grant and Refugee Rights, Inc. (ICIRR) brought one of those cases in the Northern District of Illinois. They immediately sought a preliminary injunction against the Rule pending the outcome of the litigation. Finding that the criteria for interim relief were satisfied, the district court granted their motion. We conclude that at least Cook County adequately estab- lished its right to bring its claim and that the district court did not abuse its discretion by granting preliminary injunctive re- lief. We therefore affirm. I. The Setting A. The Public-Charge Rule The Immigration and Nationality Act (INA, or “the Act”) provides that a noncitizen may be denied admission or ad- No. 19-3169 3 justment of status if she “is likely at any time to become a pub- lic charge.” 8 U.S.C. § 1182(a)(4)(A). The statute does not de- fine the term “public charge,” nor has it ever done so. Instead, the Act calls for a “totality-of-the-circumstances” analysis, though it singles out several factors to be considered “at a minimum”: age; health; family status; assets, resources, and financial status; education and skills; and any affidavit of sup- port under section 1183a. Id. § 1182(a)(4)(B). The statute does not specify how officials should weigh the listed factors and any others that appear to be relevant. On August 14, 2019, following a notice and comment pe- riod, DHS issued a rule interpreting this provision. In it, DHS defines as a “public charge” any noncitizen (with some excep- tions) who receives certain cash and noncash government benefits for more than “12 months” in the aggregate in a 36-month period. Inadmissibility on Public Charge Grounds, 84 Fed. Reg. 41292–508 (Aug. 14, 2019) (“Rule”). It applies to all legally admitted immigrants; we are not concerned here with those in the country unlawfully. The Rule is not limited to federal benefits; instead, it sweeps in any federal, state, lo- cal, or tribal cash assistance for income maintenance; Supple- mental Nutrition Assistance Program (SNAP) benefits; most forms of Medicaid; Section 8 Housing Assistance under the Housing Choice Voucher Program; Section 8 Project-Based Rental Assistance; and certain other forms of subsidized housing. Id. at 41295, 41501. Each benefit received, no matter how small, is counted separately and stacked, such that re- ceipt of multiple benefits in one month is considered receipt of multiple months’ worth of benefits. Id. at 41295. For exam- ple, an immigrant who receives any amount of SNAP benefits, Medicaid, and housing assistance, and nothing else for four months in a three-year period, will be considered a public 4 No. 19-3169 charge and likely denied adjustment of status. The stacking rule means that a person can use up her “12 months” of ben- efits in a far shorter time than a quick reading of the Rule would indicate. The Rule also explains what facts DHS will consider with respect to an applicant’s age, health, family status, financial status, and education and skills. Id. at 41502–04. “Heavily weighted negative factors” include the following: lack of cur- rent employment or reasonable prospect of future employ- ment; previous receipt or approval for receipt of 12 months’ worth of public benefits in a three-year period; diagnosis of a medical condition that is likely to require extensive medical treatment or institutionalization or that will interfere with the ability to provide for oneself, attend school, or work, along with lack of insurance and no prospect of obtaining private health insurance, and insufficient financial resources to pay for reasonably foreseeable medical costs related to such med- ical condition; and prior determination of inadmissibility or deportability on public-charge grounds. Id. at 41504. The “heavily weighted positive factors” are exclusively monetary. They include the following: a household income, assets, resources, or support amounting to at least 250 percent of the Federal Poverty Guidelines for the household size; cur- rent employment with an annual income of at least 250 per- cent of the Federal Poverty Guidelines for the household size; and private health insurance other than subsidized insurance under the Affordable Care Act. Id. To put this in perspective, recall that the Federal Poverty Guideline in 2020 for a family of four is $26,200 in annual income. Poverty Guidelines, www.aspe.hhs.gov. An annual income 250 percent of that No. 19-3169 5 number is $65,500, which is very close to the median U.S. in- come of $63,179 (the 2018 number reported by the U.S. Census on Sept. 10, 2019, see Income, Poverty and Health Insurance Coverage in the United States: 2018, www.census.gov). Other factors include whether an immigrant is younger than 18 or older than 61 (bad); household size (smaller is bet- ter); whether an immigrant’s household annual gross income is at least 125 percent of the Federal Poverty Guidelines; past receipt of any amount of public benefits (bad); level of educa- tion (good); English language proficiency; and credit history and credit score. Id. at 41502–04. The Rule represents a striking departure from the previ- ous administrative guidance—one with a potentially devast- ing impact on those to whom it applies. 1 That guidance, is- sued in 1999 by the Immigration and Naturalization Service (the predecessor of today’s U.S. Citizenship and Immigration Services), defines as a public charge a noncitizen who is “pri- marily dependent on the government for subsistence, as demonstrated by either (i) the receipt of public cash assistance for income maintenance or (ii) institutionalization for long- 1 The dissent emphasizes the fact that the Rule will not affect certain people, such as those for whom a sponsor has furnished an affidavit of support. But those are not the people who concern Cook County—it must deal with those who bear the brunt of the Rule. Cf. Planned Parenthood of Southeastern Pennsylvania v. Casey, 505 U.S. 833, 894 (1992) (“Legislation is measured for consistency with the Constitution by its impact on those whose conduct it affects. … The proper focus of constitutional inquiry is the group for whom the law is a restriction, not the group for whom the law is irrelevant.”). The dissent concedes, as it must, that the affected group is not the null set. 6 No. 19-3169 term care at government expense.” Field Guidance on Deport- ability and Inadmissibility on Public Charge Grounds, 64 Fed. Reg. 28689 (May 26, 1999) (“1999 Field Guidance”) (emphasis added); see also Proposed Rule: Inadmissibility and Deporta- bility on Public Charge Grounds, 64 Fed. Reg. 28676 (May 26, 1999). Drawing on both dictionary definitions and the devel- opment of immigration law since the 1880s, the proposed rule accompanying the 1999 Field Guidance explained that “a per- son becomes a public charge when he or she is committed to the care, custody, management, or support of the public,” and that the term is best understood to signify “a complete, or nearly complete, dependence on the Government rather than the mere receipt of some lesser level of financial support.” 64 Fed. Reg. at 28677. B. Procedural History DHS’s new rule was scheduled to go into effect in October 2019. Before it did so, plaintiffs filed this suit against DHS and related entities for declaratory and injunctive relief. The com- plaint presents several theories: (1) the Rule violates the Ad- ministrative Procedure Act (APA), 5 U.S.C. § 706, because it exceeds DHS’s statutory authority; (2) the Rule violates APA section 706 because it is not in accordance with law; (3) the Rule violates APA section 706 because it is arbitrary and ca- pricious; and (4) the Rule violates the Fifth Amendment’s Equal Protection guarantee because it discriminates against non-white immigrants. Focusing on the APA theories, plaintiffs moved for a pre- liminary injunction, which the district court granted on Octo- ber 14, 2019. (Following plaintiffs’ lead, we do not discuss the Equal Protection theory.) The injunction is geographically limited to Illinois. The district court concluded that both Cook No. 19-3169 7 County and ICIRR have constitutional standing to sue—Cook County primarily because of the added costs its health and hospital system is absorbing and will have to absorb as a re- sult of decreased immigrant enrollment in government-pro- vided health care coverage, and ICIRR because it is expending and will continue to expend additional resources to educate immigrant communities about the Rule and ensure they are able to obtain necessary health services. The court also deter- mined that both the County and ICIRR fall within the “zone of interests” protected by the INA, for largely the same rea- sons they have constitutional standing. On the merits, the court concluded that DHS’s reinterpretation of the term is likely impermissible. The court found the statute to be clear and to require more substantial, sustained dependence on government assistance than the Rule demands before a noncitizen may be considered a public charge. This showed, the court held, that plaintiffs are likely to succeed on their claims. Finally, the court ruled that plaintiffs had shown a likelihood of irreparable harm and that the balance of harms favored them, such that a preliminary injunction is war- ranted. DHS filed an immediate appeal and moved to stay the pre- liminary injunction pending resolution of its appeal. We de- nied the stay and a renewed motion for a stay, but on Febru- ary 21, 2020, the Supreme Court granted a stay. Chad Wolf, et al. v. Cook County, et al., 140 S. Ct. 681 (2020). As we write, parallel cases are being litigated in New York, Maryland, California, and Washington. New York v. U.S. Dep’t of Homeland Sec., No. 19-cv-7777 (S.D.N.Y.); Make the Road New York v. Cuccinelli, No. 19-cv-7993 (S.D.N.Y.); Casa de Maryland, Inc. v. Trump, No. 19-cv-2715 (D. Md.); California v. 8 No. 19-3169 U.S. Dep’t of Homeland Sec., No. 19-cv-4975 (N.D. Cal.); La Clinica De La Raza v. Trump, No. 19-cv-4980 (N.D. Cal.); Wash- ington v. U.S. Dep’t of Homeland Sec., No. 19-cv-5210 (E.D. Wash.). The district courts in each of those cases also issued preliminary injunctions, though with nationwide effect. DHS appealed the preliminary injunctions and requested stays pending appeal. The Ninth and Fourth Circuits granted DHS’s stay requests. City and Cnty. of San Francisco v. U.S. Cit- izenship & Immigration Servs., 944 F.3d 773 (9th Cir. 2019); Casa de Maryland, Inc. v. Trump, No. 19-2222 (4th Cir. Dec. 9, 2019). The Second Circuit declined to issue a stay, but the Supreme Court granted one pending further proceedings. Dep’t of Homeland Sec., et al., v. New York, et al., 140 S. Ct. 599 (2020). Rather than discussing these opinions point-by-point, we think it better to spell out our own analysis of these issues. II. Right To Sue Plaintiffs invoked the district court’s jurisdiction under 28 U.S.C. § 1331 for their claims under the APA. DHS responds that they lack standing to sue under Article III of the Consti- tution. The district court rejected that argument. It also con- cluded that plaintiffs had adequately raised a claim within the “zone of interests” of the INA. We review the legal question of standing de novo and the factual findings underlying the district court’s determination of standing for clear error. Arre- ola v. Godinez, 546 F.3d 788, 794 (7th Cir. 2008). A. Article III Standing Article III of the Constitution limits the federal judicial power to the adjudication of “cases” and “controversies.” U.S. Const. art. III, § 2, cl. 1. For there to be a justiciable case or controversy, the party invoking the power of the court must No. 19-3169 9 have standing to sue. Hollingsworth v. Perry, 570 U.S. 693, 700 (2013). To assert standing for injunctive relief, a plaintiff must show that it is under an actual or imminent threat of suffering a concrete and particularized injury-in-fact; that this injury is fairly traceable to the defendant’s conduct; and that it is likely that a favorable judicial decision will prevent or redress the injury. Summers v. Earth Island Inst., 555 U.S. 488, 493 (2009). Municipalities generally have standing to challenge laws that result (or immediately threaten to result) in substantial financial burdens and other concrete harms. See, e.g., Dep’t of Commerce v. New York, 139 S. Ct. 2551, 2565 (2019) (“diminish- ment of political representation, loss of federal funds, degra- dation of census data, and diversion of resources” were suffi- cient to give states and municipalities standing to sue over the proposed inclusion of a citizenship question on the 2020 cen- sus); Gladstone, Realtors v. Vill. of Bellwood, 441 U.S. 91, 110–11 (1979) (municipality had standing based on the effect of racial steering in housing on the municipality’s tax base and social stability). The district court found that Cook County has standing based on the financial harms the County will incur if and when the Rule goes into effect. The Rule is likely to cause im- migrants to forgo routine treatment, immunizations, and di- agnostic testing, resulting in more costly, uncompensated emergency care and an increased risk of communicable dis- eases spreading to the general public. Indeed, DHS conceded this harm in its commentary on the Rule, acknowledging “that increased use of emergency rooms and emergent care as a method of primary healthcare due to delayed treatment is possible and there is a potential for increases in uncompen- sated care in which a treatment or service is not paid for by an 10 No. 19-3169 insurer or patient.” 84 Fed. Reg. at 41384. The district court determined that “[b]oth the costs of community health epi- demics and of uncompensated care are likely to fall particu- larly hard on [the Cook County health system], which already provides approximately half of all charity care in Cook County, including to noncitizens regardless of their immigra- tion status.” The district court found that these financial and health burdens were sufficient. The district court also concluded that ICIRR has Article III standing based on the effect of the Rule on its ability to per- form its core mission and operate its existing programs. The court found that the Rule would impair the organization’s ability to achieve its mission of increasing access to care, im- proving health literacy, and reducing reliance on emergency room care in immigrant communities. The Rule already has caused ICIRR to divert resources from its core programs to new efforts designed to educate immigrants and staff about the Rule’s effects and to mitigate the Rule’s chilling impact on immigrants who are not covered by the Rule but who none- theless fear immigration consequences based on their receipt of public benefits. Under Havens Realty Corp. v. Coleman, 455 U.S. 363 (1982) and Common Cause Indiana v. Lawson, 937 F.3d 944 (7th Cir. 2019), this is enough. In Havens, the Supreme Court found that a nonprofit organization focused on equal housing access had standing to sue an apartment owner under the Fair Housing Act for racial discrimination, based on the negative impact of the defendant’s racial steering practices on the organization’s ability to provide counseling and referral services for low-and moderate-income home-seekers. 455 U.S. at 379. And in Com- mon Cause, we relied on Havens in concluding that the plaintiff No. 19-3169 11 voting rights organizations had standing to challenge an In- diana law designed to remove certain people from the voter rolls, because the law caused the organizations to divert their limited resources from core programs to ameliorating the ef- fects of the law. 937 F.3d at 950–52. We agree with the district court that Cook County and ICIRR have established cognizable injuries. Their alleged in- juries are predictable, likely, and imminent. And the Rule— not independent third-party decision-making—is the but-for cause of these injuries. Plaintiffs thus have constitutional standing to challenge the Rule. B. Statutory Coverage The next question is whether the interests Cook County and ICIRR assert are among those protected or regulated by the INA. A statute “ordinarily provides a cause of action only to plaintiffs whose interests fall within the zone of interests protected by the law invoked.” Bank of America Corp. v. City of Miami, 137 S. Ct. 1296, 1302 (2017). The zone-of-interests test is not “especially demanding” in the APA context. Lexmark Int’l, Inc. v. Static Control Compo- nents, Inc., 572 U.S. 118, 130 (2014). This is because it was “Congress’s evident intent when enacting the APA to make agency action presumptively reviewable.” Match-E-Be-Nash- She-Wish Band of Pottawatomi Indians v. Patchak, 567 U.S. 209, 225 (2012). The plaintiffs’ interests must only arguably fall within the zone of interests of the statute. And the emphasis on the word “arguably” is not ours: the Supreme Court has “always conspicuously included the word ‘arguably’ in the test to indicate that the benefit of any doubt goes to the plain- tiff.” Id. It is not necessary to demonstrate any “indication of 12 No. 19-3169 congressional purpose to benefit the would-be plaintiff.” Id. Suit is foreclosed “only when a plaintiff’s interests are so mar- ginally related to or inconsistent with the purposes implicit in the statute that it cannot reasonably be assumed that Con- gress intended to permit the suit.” Id. 1. Cook County The district court concluded that Cook County satisfies the zone-of-interests test based on the financial burdens the County will incur as a result of the Rule. It drew an analogy to City of Miami, in which the Supreme Court held that Mi- ami’s allegations of lost tax revenue and extra municipal ex- penses placed it within the zone of interests protected by the Fair Housing Act. DHS takes issue with these conclusions. It argues that the County does not fall within the INA’s zone of interests be- cause its asserted interests are inconsistent with the statutory purpose. DHS sees a tension between the County’s efforts to provide services to immigrants and the supposed aim of the public-charge provision, which it understands as a command to reduce and penalize immigrants’ receipt of public benefits. DHS also contends that the district court misread City of Mi- ami, and that the INA does not give any third party a judicially enforceable interest in the Executive Branch’s immigration decisions. DHS has overshot the mark. Indeed, its own arguments undermine such an absolutist position. DHS admits that one purpose of the public-charge provision is to protect taxpayer resources. In large measure, that is the same interest Cook County asserts. DHS tries to distinguish itself from Cook County by saying that it is focused on reducing the burden on No. 19-3169 13 federal taxpayers, but the Rule itself covers not just federal, but also state, local, and tribal assistance. Even if the effect of the Rule is some reduction in the burden on federal taxpayers, Cook County has plausibly alleged that at the same time, the Rule will increase the burden on those same people in their capacity as state and local taxpayers, who will have to suffer the adverse effects of a substantial population with inade- quate medical care, housing, and nutrition. Furthermore, though the purpose of the public-charge provision is to screen for and promote “self-sufficiency” among immigrants, it is not obvious what self-sufficiency means. Subsidies abound in the modern world, from dis- counted or free transportation for seniors, to public snow re- moval, to school lunches, to childhood vaccinations, and much more. Cf. Danilo Trisi, Administration’s Public Charge Rules Would Close the Door to U.S. Immigrants Without Substan- tial Means, Ctr. on Budget and Policy Priorities (Nov. 11, 2019) (noting that in a single year, one in four U.S.-born citizens, and 15 percent of all residents, receives a benefit included in the Rule’s public charge definition). Ensuring that immigrants have access to affordable basic health care, for example, may promote their greater self-sufficiency in other domains, in- cluding income, housing, and nutrition. It also protects the community at large from highly contagious diseases such as COVID-19. Cook County’s interest in ensuring lawful immi- grants’ access to authorized federal and state public benefits is not plainly inconsistent with the text of the statute. Its fi- nancial interests thus suffice to bring it within the zone of in- terests of the public-charge provision. 14 No. 19-3169 2. ICIRR The court also found that ICIRR fits within the INA’s zone of interests, explaining that there is “ample evidence that ICIRR’s interests are not merely marginal to those of the aliens more directly impacted by the public charge provision” and that “ICIRR [is] precisely the type of organization that would reasonably be expected to ‘police the interests that the statute protects.’” Because only one plaintiff need demonstrate that it has stated a claim within the zone of interests of the statute, we elect to pass over ICIRR without much comment. We recog- nize that it asserts that it has suffered a financial burden di- rectly attributable to the Rule. And we accept that ICIRR helps immigrants navigate the INA’s various requirements, includ- ing the public-charge rule, and it has an interest in ensuring that immigrants are not improperly denied adjustment of sta- tus or removed from the country because of confusion over DHS’s Rule. But the link between these injuries and the pur- pose of the public-charge part of the statute is more attenu- ated, and thus it is harder to say that the injury ICIRR has as- serted meets the “zone-of-interests” test. Given Cook County’s presence in the case, we need not resolve ICIRR’s status definitively, and so we limit our discus- sion in the remainder of the opinion to Cook County. The cen- tral question is whether the district court abused its discretion in preliminarily enjoining the Rule for the State of Illinois? III. The Preliminary Injunction To obtain a preliminary injunction, a plaintiff must estab- lish that: (1) she is likely to succeed on the merits, (2) she is likely to suffer irreparable harm in the absence of preliminary No. 19-3169 15 relief; and (3) legal remedies are inadequate. See Winter v. Nat- ural Res. Def. Council, Inc., 555 U.S. 7, 20 (2008); Eli Lilly & Co. v. Arla Foods, Inc., 893 F.3d 375, 381 (7th Cir. 2018). “If the mov- ing party makes this showing, the court balances the harms to the moving party, other parties, and the public.” Eli Lilly, 893 F.3d at 381. The standard is the same for an application for a stay under section 705 of the APA. Cronin v. U.S. Dep’t of Agric., 919 F.2d 439, 446 (7th Cir. 1990). The district court concluded that Cook County is likely to succeed on the merits and that the other requirements for pre- liminary injunctive relief have been met. We review the issu- ance of a preliminary injunction under the deferential abuse- of-discretion standard, reviewing legal issues de novo and fac- tual findings for clear error. Whitaker v. Kenosha Unified Sch. Dist. No. 1 Bd. of Educ., 858 F.3d 1034, 1044 (7th Cir. 2017). A. Likelihood of Success The pivotal question in this case, as in many involving pre- liminary relief, is likelihood of success on the merits. We therefore devote the bulk of our analysis to this issue, under- standing that the litigation is still in an early stage and any- thing we say may change as the record develops further. The APA provides for judicial review of final agency deci- sions. 5 U.S.C. §§ 702, 706. The overriding question is whether the agency’s interpretation of the relevant statute is one the text will permit. We approach this inquiry through the two- step framework set forth in Chevron U.S.A. Inc. v. Nat. Res. Def. Council, Inc., 467 U.S. 837 (1984). The first issue is “whether Congress has directly spoken to the precise question at issue.” Id. at 842. If Congress has done so unambiguously, then that is the end of it: the agency and courts alike are bound by what 16 No. 19-3169 Congress wrote. Id. at 842–43. If Congress has not spoken clearly, then we move on to step two, in which we consider whether the agency’s interpretation reflects a permissible con- struction of the statute. Id. at 843. We defer to the agency’s reading “unless it appears from the statute or its legislative history that the accommodation [of conflicting policies] is not one that Congress would have sanctioned.” Id. at 845; see also Indiana v. EPA, 796 F.3d 803, 811 (7th Cir. 2015). Statutory interpretation is not the end of the matter, how- ever. We also must assess the agency’s policymaking to en- sure that it is not “arbitrary and capricious,” as the APA uses those terms. 5 U.S.C. § 706(2)(A). This review, guided by Mo- tor Vehicle Mfrs. Ass’n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29 (1983), focuses not on the facial validity of the agency’s in- terpretation, but rather on the soundness of the process by which it reached its interpretation. See Encino Motorcars, LLC v. Navarro, 136 S. Ct. 2117, 2125 (2016) (“[W]here a proper challenge is raised to the agency procedures, and those pro- cedures are defective, a court should not accord Chevron def- erence to the agency interpretation.”). 1. Chevron Step One We begin our analysis of DHS’s Rule with an analysis of the text of the INA. In conducting this analysis, we consider the words of the public-charge provision, its place in the over- all statutory scheme, the relation of the INA to other statutes, and “common sense as to the manner in which Congress is likely to delegate a policy decision of such economic and po- litical magnitude to an administrative agency.” FDA v. Brown & Williamson Tobacco Corp., 529 U.S. 120, 132–33 (2000). No. 19-3169 17 As we noted at the outset, the INA contains no formal def- inition of what it takes to be a “public charge.” It merely lists several broad factors that are relevant to the determination. 8 U.S.C. § 1182(a)(4)(B). It does not provide weights for either the listed factors or any others that might exist in a given case. Instead, it relies on the discretion of the responsible consular official or the Attorney General. Id. § 1182(a)(4)(A). It also au- thorizes the Secretary of Homeland Security to promulgate rules to guide those determinations. Id. § 1103(a)(3). In defense of its Rule, DHS relies heavily on the 1996 amendments to the INA. There Congress stated that “self-suf- ficiency has been a basic principle of United States immigra- tion law since this country’s earliest immigration statutes.” 8 U.S.C. § 1601(1). Congress also announced its intent that “al- iens within the Nation’s borders not depend on public re- sources to meet their needs, but rather rely on their own ca- pabilities and the resources of their families, their sponsors and private organizations”; and that “the availability of pub- lic benefits [should] not constitute an incentive for immigra- tion to the United States.” 8 U.S.C. § 1601(2). Both parties also cite various other statutory provisions that they believe shed light on the meaning of the public- charge provision, including the requirement for some immi- grants to obtain affidavits of support from sponsors, 8 U.S.C. § 1183a; an exception to the public-charge provision for immi- grants who are victims of domestic violence and receive ben- efits in that capacity, id. §§ 1182(a)(4)(E), 1641(c); and several other statutes that extend, with varying conditions, certain benefits to immigrants, see, e.g., id. §§ 1611, 1621. We do not find these provisions to be particularly helpful; each is sus- ceptible to more than one reasonable reading. 18 No. 19-3169 Cook County argues that long-established judicial deci- sions, ratified by Congress, point us to only one possible in- terpretation—that is, the one that it urges. But in our view, the historical record is not so clear. The parties agreed in the dis- trict court that the understanding of the term “public charge” around the time it first entered federal immigration law in 1882 is particularly relevant. See New Prime Inc. v. Oliveira, 139 S. Ct. 532, 539 (2019) (“It’s a fundamental canon of statutory construction that words generally should be interpreted as taking their ordinary meaning at the time Congress enacted the statute.”). But that is where the harmony ends. Enter the dueling dictionaries. In Cook County’s corner, we have the Century Dictionary, defining a “charge” as a per- son who is “committed to another’s custody, care, concern or management,” Century Dictionary 929 (William Dwight Whitney, ed., 1889) (emphasis added); and Webster’s Diction- ary, likewise defining a “charge” as a “person or thing com- mitted to the care or management of another,” Webster’s Con- densed Dictionary of the English Language 84 (Dorsey Gard- ner, ed., 1884). These suggest primary, long-term dependence. In DHS’s corner, we have dictionaries defining a “charge” as “an obligation or liability,” as in a “pauper being chargeable to the parish or town,” Dictionary of Am. and English Law 196 (Stewart Rapalje & Robert Lawrence, eds., 1888); and as a “burden, incumbrance, or lien,” Glossary of the Common Law 56 (Frederic Jesup Stimson, ed., 1881). These definitions can be read to indicate that a lesser reliance on public benefits is enough. Finding no clarity here, we move on. Cook County contends that from the outset Congress dis- tinguished between, on the one hand, those who were permit- No. 19-3169 19 ted to “land” and receive short-term support from govern- ment agencies, and, on the other hand, those who were ex- cluded as public charges. Under the 1882 Immigration Act, the set of people who could be prevented from landing in- cluded convicts, “lunatics,” “idiots,” and any other person “unable to take care of himself or herself.” An Act to Regulate Immigration, ch. 376, § 2, 22 Stat. 214 (1882). The 1882 Act au- thorized the Secretary of the Treasury, who was responsible for supervising immigration, to enter into contracts with state entities “to provide for the support and relief of such immi- grants therein landing as may fall into distress or need public aid, under the rules and regulations to be prescribed by said Secretary.” Id. Cook County stresses this distinction between excludable public charges and immigrants who (less drasti- cally) “may fall into distress or need public aid.” This argument has some intuitive merit. DHS responds however, that the general revenues were not at risk under the 1882 Act for immigrants who were not self-sufficient upon ar- rival. The 1882 Act directed the Secretary of the Treasury to levy an entry tax on all noncitizens arriving by ship to cover both the cost of regulating immigration and that of temporary assistance. 1882 Act, ch. 376, § 1. It also specified that “no greater sum shall be expended for the purposes hereinbefore mentioned, at any port, than shall have been collected at such port.” Id. In other words, federal funding was available only to the extent the funds matched collections from the vessels. This general feature is no longer part of the law (putting to one side the special case of sponsored immigrants). Congress tinkered with the language in 1891, 1903, and 1907. See An Act in Amendment to the Various Acts Relative to Immigration and the Importation of Aliens Under Contract 20 No. 19-3169 or Agreement to Perform Labor, ch. 551, 26 Stat. 1084, 1086 (1891); An Act to Regulate the Immigration of Aliens into the United States, ch. 1012, 32 Stat. 1213, 1213–1214 (1903); An Act to Regulate the Immigration of Aliens into the United States, ch. 1134, 34 Stat. 898, 898–899, 904–905 (1907). Never, how- ever, did it define “public charge” or explain what degree of reliance on government aid brands someone as such a person. Federal district court and state-court cases from this pe- riod point in different directions. For example, in In re Fein- knopf, 47 F. 447 (E.D.N.Y. 1891), a district court distinguished between the primary dependence of persons who live in alms- houses and the lesser dependence of those who merely re- ceive public support. Around the same time, a North Dakota court indicated that temporary aid is actually a means of averting public dependence, insofar as it can keep those “des- titute of means and credit from becoming a public charge.” Yeatman v. King, 2 N.D. 421 (1892). On the other hand, the question in Yeatman was whether an obligation to repay the county the value of received temporary public assistance counted as a tax, and so the decision is of limited value. The district court did not find these and other early deci- sions to be dispositive. Instead, it thought that the Supreme Court’s decision in Gegiow v. Uhl, 239 U.S. 3 (1915), resolved the issue. But there, too, the question presented was a narrow one. The Court said that it was addressing the “single ques- tion … whether an alien can be declared likely to become a public charge on the ground that the labor market in the city of his immediate destination is overstocked.” Id. at 9–10. It an- swered in the negative, saying that “[t]he persons enumer- ated, in short, are to be excluded on the ground of permanent personal objections accompanying them irrespective of local No. 19-3169 21 conditions … .” Id. at 10. The district court in our case under- stood Gegiow as holding that the term “public charge” encom- passes only persons who are substantially, if not entirely, de- pendent on government assistance on a long-term basis. While there is language in Gegiow that supports that read- ing, we are not persuaded that the Supreme Court necessarily ruled so broadly. The Court went out of its way to say that the question presented was the one we noted above. The Acting Commissioner of Immigration, in deciding to deport the per- sons at issue, mentioned in addition to local labor conditions “the amount of money possessed and ignorance of our lan- guage.” But the Court brushed off these considerations as mere “makeweights.” Id. at 9. It thus had no need to address directly the immigrants’ financial resources and education. In context, the Court’s reference to “permanent personal objections” might have simply reflected a distinction between the individualized characteristics of an immigrant and exter- nal factors such as a local labor market. The terse opinion is silent about any distinction between people whose need for public assistance is temporary and minimal, and those whose need is likely to be substantial or permanent. We thus agree with DHS that the case before us cannot be resolved exclu- sively by reference to Gegiow. Circuit court decisions in the aftermath of Gegiow add little clarity to this picture. For example, in Wallis v. United States ex rel. Mannara, 273 F. 509 (2d Cir. 1921), the Second Circuit de- fined a person likely to become a public charge as “one whom it may be necessary to support at public expense by reason of poverty, insanity and poverty, disease and poverty, idiocy and poverty.” Id. at 511. It did so in a case in which the immi- 22 No. 19-3169 grant family’s primary breadwinner was “certified for senil- ity” and thus would never be “capable of continued self-sup- port.” Id. at 510. The court noted that the family had “insuffi- cient [means] to provide for their necessary wants any reason- able length of time” and no private sources of support. Id. On the other hand, in Ex parte Hosaye Sakaguchi, 277 F. 913 (9th Cir. 1922), the Ninth Circuit held that an immigrant woman with the skills to support herself was not likely to become a public charge. Id. at 916. It ruled that the government had to present evidence of “mental or physical disability or any fact tending to show that the burden of supporting the [immi- grant] is likely to be cast upon the public.” Id. How much of a burden was left undefined. See also United States ex rel. De Sousa v. Day, 22 F.2d 472, 473–74 (2d Cir. 1927) (“In the face of [Gegiow] it is hard to say that a healthy adult immigrant, with no previous history of pauperism, and nothing to interfere with his chances in life but lack of savings, is likely to become a public charge within the meaning of the statute.”). The parties and amici also call our attention to later actions by the Executive Branch, but we find these also to be incon- clusive. See, e.g., Matter of B-, 3 I. & N. Dec. 323, 326 (BIA & AG 1948) (stating that the longstanding test for whether an immigrant could be deemed a public charge had three com- ponents: (1) the state must charge for the service it renders; (2) it must make a demand for payment; and (3) the immigrant must fail to pay). What we can say is that in 1952 Congress amended the Act in a way that uses the language of discretion: it deems inad- missible immigrants “who, in the opinion of the consular of- ficer at the time of application for a visa, or in the opinion of the Attorney General at the time of application for admission, No. 19-3169 23 are likely at any time to become public charges.” An Act to Re- vise the Laws Relating to Immigration, Naturalization, and Nationality; and for Other Purposes, Pub. L. No. 414, § 212, 66 Stat. 163, 183 (1952) (emphasis added). This language clarifies the temporal dimension of the public-charge determination, but it says nothing about the degree or duration of assistance. The Immigration Act of 1990, Pub. L. No. 101-649, § 601, 104 Stat. 4978, also lacks a clear definition of “public charge.” In the 1996 Immigration Act, Congress for the first time provided guidance on what the Executive Branch must con- sider when determining whether an immigrant is likely to be- come a public charge. As we noted earlier, immigration offi- cials were instructed “at a minimum” to look at age, health, family status, financial situation, and education and skills. Il- legal Immigration Reform and Immigrant Responsibility Act of 1996, Pub. L. No. 104-208, Div. C, § 531, 110 Stat. 3009 (1996). They also could consider whether an immigrant had an affidavit of support from a third party. Id. Congress re- jected a proposal to define “public charge” to cover “any alien who receives [means-tested public benefits] for an aggregate of at least 12 months.” 142 Cong. Rec. 24313, 24425 (1996). Contemporaneously, Congress enacted the Personal Re- sponsibility and Work Opportunity Reconciliation Act of 1996, Pub. L. 104-193, 110 Stat. 2105 (1996), commonly known as the “Welfare Reform Act.” DHS places great weight on lan- guage in that statute’s expression of Congress’s desire that immigrants be self-sufficient and not come to the United States with the purpose of benefitting from public welfare programs. See 8 U.S.C. § 1601(1). The INA (with that amend- ment) pursues that goal by restricting most noncitizens from eligibility for many federal and state public benefits. It grants 24 No. 19-3169 lawful permanent residents access to means-tested public benefits only after they have spent five years as a lawful per- manent resident. Id. §§ 1611, 1613, 1621. But the exclusions are not absolute. Congress specified instead that immigrants may at any time receive emergency medical assistance; immuniza- tions and testing for communicable diseases; short-term, in- kind emergency disaster relief; various in-kind services such as short-term shelter and crisis counseling; and certain hous- ing and community development assistance. Id. The INS summarized its understanding of the 1996 legal regime in the 1999 Field Guidance, which defined as a public charge those who are “primarily dependent on the govern- ment for subsistence, as demonstrated by either (i) the receipt of public cash assistance for income maintenance or (ii) insti- tutionalization for long-term care at government expense.” 64 Fed. Reg. at 28689. Following an earlier 1987 interpretive rule, see Adjustment of Status for Certain Aliens, 52 Fed. Reg. 16205, 16211–12, 16216 (May 1, 1987), the 1999 Field Guidance said that “officers should not initiate or pursue public charge deportation cases against aliens who have not received public cash benefits for income maintenance or who have not been institutionalized for long-term care.“ 64 Fed. Reg. at 28689. It directed officers “not [to] place any weight on the receipt of non-cash public benefits (other than institutionalization) or the receipt of cash benefits for purposes other than for income maintenance with respect to determinations of admissibility or eligibility for adjustment on public charge grounds.” Id. Later enactments lightened some of the statutory re- strictions, in order to allow additional categories of immi- grants to qualify for certain benefits without a five-year wait- ing period. See Farm Security and Rural Investment Act of No. 19-3169 25 2002, Pub. L. 107-171, § 4401, 116 Stat. 134 (2002); Children’s Health Insurance Program Reauthorization Act of 2009, Pub. L. 11-3, § 214, 123 Stat. 8 (2009). This is where things stood when DHS developed the Rule. What should we make of this historical record? As the district court recognized, there is abundant evidence supporting Cook County’s interpretation of the public-charge provision as being triggered only by long-term, primary dependence. But the question before us is not whether Cook County has offered a reasonable interpretation of the law. It is whether the statutory language unambiguously leads us to that interpre- tation. We cannot say that it does. As our quick and admit- tedly incomplete overview of this byzantine law has shown, the meaning of “public charge” has evolved over time as im- migration priorities have changed and as the nature of public assistance has shifted from institutionalization of the destitute and sick, to a wide variety of cash and in-kind welfare pro- grams. What has been consistent is the delegation from Con- gress to the Executive Branch of discretion, within bounds, to make public-charge determinations. Thus, this case cannot be resolved at Chevron step one. But that does not end the analysis, because we may affirm the dis- trict court’s issuance of a preliminary injunction on any basis in the record. See Valencia v. City of Springfield, Ill., 883 F.3d 959, 967 (7th Cir. 2018). We therefore proceed to step two. 2. Chevron Step Two At step two of the Chevron analysis, we consider “whether the agency’s answer is based on a permissible construction of the statute.” 467 U.S. at 843. Our review is deferential; we ac- cord “considerable weight … to an executive department’s 26 No. 19-3169 construction of a statutory scheme it is entrusted to adminis- ter.” Id. at 844; see also Nat’l Cable & Telecomms. Ass’n v. Brand X Internet Servs., 545 U.S. 967, 980 (2005). A court may strike down an agency’s interpretation of a law if, for example, the agency’s reading disregards the stat- utory context, see, e.g., Michigan v. EPA, 135 S. Ct. 2699, 2708 (2015); its rule is based on an unreasonable interpretation of legislative history, see, e.g., Council for Urological Interests v. Burwell, 790 F.3d 212, 223 (D.C. Cir. 2015); or its new position “would bring about an enormous and transformative expan- sion in [the agency’s] regulatory authority without clear con- gressional authorization, Util. Air Reg. Grp. v. EPA, 573 U.S. 302, 324 (2014). Cook County offers several reasons why DHS’s interpre- tation founders here. First, it contends that the Rule conflicts with at least two statutes: the SNAP statute and the Rehabili- tation Act of 1973. Second, it urges that the DHS position cre- ates internal inconsistencies in the immigration laws them- selves. We address these points in turn. The SNAP statute prohibits the government from consid- ering SNAP benefits as “income or resources for any purpose under any Federal, State, or local laws.” 7 U.S.C. § 2017(b). But DHS is not trying to characterize these benefits as income or resources held by the immigrant in question. The Rule merely notes that receipt of the benefits is an indicium of a lack of self-sufficiency. Whatever else one might say about that posi- tion, it is not one that the SNAP law forbids. The Rehabilitation Act of 1973 prohibits the government from excluding from participation in, denying the benefits of, or subjecting to discrimination under any federally funded No. 19-3169 27 program or activity, a person with a disability “solely by rea- son of her or his disability.” 29 U.S.C. § 794(a). An agency vi- olates the Act if it (1) intentionally acts on the basis of the dis- ability; (2) refuses to provide a reasonable modification; or (3) takes an action or adopts a rule that disproportionately affects disabled people. A.H. ex rel. Holzmueller v. Ill. High Sch. Ass’n, 881 F.3d 587, 592–93 (7th Cir. 2018). An aggrieved person must demonstrate that “but for” her disability, she would have been able to access the desired benefits. Id. at 593. DHS frankly acknowledges that it takes disability into ac- count in its public-charge analysis, and it does so in an unfa- vorable way. 84 Fed. Reg. at 41383 (“DHS considers any disa- bility or other medical condition in the public charge inadmis- sibility determination to the extent the alien’s health makes the alien more likely than not to become a public charge at any time in the future.”). Indeed, the Rule brands as a heavily weighted negative factor a medical condition that is likely to require extensive medical treatment or interfere with the per- son’s ability to provide for herself, attend school, or work. Id. at 41504. DHS does not say what amounts to “extensive med- ical treatment” or what it means for a condition to “interfere with [an immigrant’s] ability to provide for herself, attend school, or work.” The Rule leaves the interpretation of these terms to immigration officials. It is therefore unclear what sorts of disabilities DHS will place into this category. As several amici curiae point out, the Rule ignores the fact that private insurers do not cover many home- and commu- nity-based services, and so denial of benefits is effectively de- nial of access to programs or activities. See id. at 41382. DHS responded to this criticism, as it applies to Medicaid Buy-in for those with disabilities, with the comment that “[a]liens 28 No. 19-3169 should be obtaining private health insurance other than Med- icaid in order to establish self-sufficiency.” Id. But that is chi- merical. Private insurance in the United States typically ex- cludes these benefits, and so persons with disabilities are able to obtain essential services, including personal-care services, specialized therapies and treatment, habilitative and rehabil- itative services, and medical equipment, only by participating in the Medicaid Buy-in program. With this assistance, they are able to work and thus can avoid becoming a public charge, which is DHS’s purported goal. The conclusion is inescapable that the Rule penalizes dis- abled persons in contravention of the Rehabilitation Act. All else being equal—education and skills, work history and po- tential, health besides disability, etc.—the disabled are sad- dled with at least two heavily weighted negative factors di- rectly as a result of their disability. Even while DHS purports to follow the statutorily-required totality of the circumstances test, the Rule disproportionately burdens disabled people and in many instances makes it all but inevitable that a person’s disability will be the but-for cause of her being deemed likely to become a public charge. We do not mean to suggest that the Rehabilitation Act re- pealed the “health” criterion in the public-charge provision by implication. There is no need to do that, if the two statutes can be reconciled—and it is our duty to see if that can be ac- complished. See Epic Sys. Corp. v. Lewis, 138 S. Ct. 1612, 1624 (2018) (“[T]his Court is not at liberty to pick and choose among congressional enactments and must instead strive to give effect to both.”). And they can live together comfortably, as long as we understand the “health” criterion in the INA as referring to things such as contagious disease and conditions No. 19-3169 29 requiring long-term institutionalization, but not disability per se. That interpretation is also historically grounded. DHS’s interpretation also creates serious tensions, if not outright inconsistencies, within the statutory scheme. It con- flicts with Congress’s affirmative authorization for desig- nated immigrants to receive the benefits the Rule targets. See 8 U.S.C. §§ 1611, 1621 (allowing immigrants to receive emer- gency medical assistance, immunizations and contagious dis- ease testing, and some public housing assistance); Farm Secu- rity and Rural Investment Act of 2002, Pub. L. No. 107–171, § 4401, 116 Stat. 134 (authorizing supplemental nutrition ben- efits for certain categories of immigrants, and Medicaid and children’s health insurance for noncitizen children and preg- nant women). Cook County is largely correct when it accuses the Rule of “set[ting] a trap for the unwary” by penalizing people for accepting benefits Congress made available to them. Although the Rule does not punish immigrants for us- ing the designated benefits, in the sense of imposing a fine, its heavily negative consideration of such use is an even worse penalty for someone seeking a lawful path to staying in the United States. Furthermore, the preliminary injunction record shows that many immigrants are not sophisticated enough to know which benefits they may safely accept and which not. Congress drew the balance between acceptance of benefits and preference for self-sufficiency in the statutes, and it is DHS’s duty to respect that outer boundary. The Welfare Re- form Act achieved its stated goal of reducing immigrant reli- ance on public assistance by barring receipt of any benefits by some classes of noncitizens and authorizing receipt by other classes only after a five-year waiting period. The statute did not create a regime that permitted self-sufficiency to trump all 30 No. 19-3169 other goals, nor did it modify the public-charge provision to penalize receipt of non-cash as well as cash assistance. DHS is correct that its Rule is not worded as an outright prohibition against an immigrant’s receipt of benefits to which Congress has entitled him. The latter would exceed DHS’s authority. But the record before us indicates that it may have the same effect. Our concerns are heightened by the fact that DHS’s inter- pretation of its statutory authority has no natural limitation. Although it chose a rule that quantified the benefits used to 12 months’ worth over a 36-month period, nothing in its in- terpretation requires even that limit. There is nothing in the text of the statute, as DHS sees it, that would prevent the agency from imposing a zero-tolerance rule under which the receipt of even a single benefit on one occasion would result in denial of entry or adjustment of status. We see no warrant in the Act for this sweeping view. Even assuming that the term “public charge” is ambiguous and thus might encompass more than institutionalization or pri- mary, long-term dependence on cash benefits, it does violence to the English language and the statutory context to say that it covers a person who receives only de minimis benefits for a de minimis period of time. There is a floor inherent in the words “public charge,” backed up by the weight of history. The term requires a degree of dependence that goes beyond temporary receipt of supplemental in-kind benefits from any type of public agency. DHS also runs into trouble as a result of its decision to stack benefits and disregard monetary value. Under its Rule, the receipt of multiple benefits in one month, no matter how No. 19-3169 31 slight, counts as multiple months of benefits. DHS acknowl- edges that the Rule’s 12-months-in-36 tolerance would actu- ally run out in four months if an immigrant received non- emergency Medicaid, any SNAP benefit, and housing assis- tance, or even sooner if she additionally received any amount of cash income assistance through a federal, state, local, or tribal program. Paradoxically, the Rule provides no oppor- tunity for an immigrant to repay the value of the benefits re- ceived once she is back on her feet. This is another way in which it unreasonably imposes substantially disproportion- ate consequences for immigrants, compared to the supposed drain on the public fisc they cause. The ambiguity in the public-charge provision does not provide DHS unfettered discretion to redefine “public charge.” We find that the interpretation reflected in the Rule falls outside the boundaries set by the statute. 3. Arbitrary and Capricious Review Our conclusion that the Rule likely does not meet the standards of Chevron step two is enough to require us to move on to the remainder of the preliminary-injunction analysis. But even if we are wrong about step two, one more inquiry remains: whether the Rule is arbitrary and capricious, as the APA uses those terms. See 5 U.S.C. § 706(2)(A). That requires an examination of DHS’s policymaking process. When conducting rulemaking, an agency must “examine the relevant data and articulate a satisfactory explanation for its action including a rational connection between the facts found and the choice made.” Motor Vehicle Mfrs. Ass’n v. State Farm, 463 U.S. at 43. It may not “rel[y] on factors which Con- gress has not intended it to consider, entirely fail[] to consider 32 No. 19-3169 an important aspect of the problem, [or] offer[] an explanation for its decision that runs counter to the evidence before the agency, or is so implausible that it could not be ascribed to a difference in view or the product of agency expertise.” Id. Fur- thermore, when an agency changes course, as DHS did here when it adopted a radically different understanding of the term “public charge” compared to the 1999 Field Guidance, it “must show that there are good reasons for the new policy.” FCC v. Fox Television Stations, Inc., 556 U.S. 502, 515 (2009). In explaining a change in policy, “an agency must also be cogni- zant that longstanding policies may have engendered serious reliance interests that must be taken into account.” Encino Mo- torcars, 136 S. Ct. at 2126. This is because a “settled course of behavior embodies the agency’s informed judgment that, by pursuing that course, it will carry out the policies committed to it by Congress.” State Farm, 463 U.S. at 41–42. Thus, “a rea- soned explanation is needed for disregarding facts and cir- cumstances that underlay or were engendered by the prior policy.” Fox Television Stations, 556 U.S. at 516. The review called for by State Farm is narrow in scope and does not permit us to substitute our own policy judgment for that of the agency. We ask only whether the agency’s “deci- sion was based on a consideration of the relevant factors” and was not “a clear error of judgment.” 463 U.S. at 43. In response to its notice of proposed rulemaking, DHS re- ceived a whopping 266,077 comments, the vast majority of which opposed the proposed rule. In the preamble to the final rule, DHS summarized significant issues raised by the com- ments and changes it made in the final rule. We assess the va- lidity of DHS’s decision-making process based on this record. No. 19-3169 33 Cook County urges that the Rule is arbitrary and capri- cious in a number of ways: (1) DHS failed meaningfully to evaluate and address significant potential harms from the Rule, including its substantial chilling effect on immigrants not covered by the Rule; (2) DHS failed to give a logical ra- tionale for the duration-based standard; and (3) DHS added factors to the totality-of-the-circumstances analysis that are “unsupported, irrational and at odds with the Final Rule’s purported purpose.” Numerous amici underscored these points and explained how the Rule will lead to arbitrary re- sults, cause both direct and indirect economic harms, burden states and localities that have to manage fallout from the Rule, and disproportionately harm the disabled and children. We look first at DHS’s dismissal of concerns about the Rule’s chilling effect on legal immigrants and family members who fall outside its scope. DHS acknowledged a “plausible connection” between the Rule and needless disenrollment by exempt noncitizens (including refugees, asylees, and victims of domestic violence) in covered public benefits, and by cov- ered immigrants in noncovered benefit programs. 84 Fed. Reg. at 41313. DHS also said that it “appreciates … the poten- tial nexus between public benefit enrollment reduction and food insecurity, housing scarcity, public health and vaccina- tions … and increased costs to states and localities.” Id. None- theless, it brushed off these impacts as “difficult to predict” and refused to “alter this rule to account for such unwar- ranted choices.” Id. Even though these consequences are fore- seeable, the Rule does not literally compel them, and so DHS asserted that they could be addressed through additional public guidance. 34 No. 19-3169 DHS may think that these responses are unwarranted, but it does not deny that they are taking place and will continue to do so. Moreover, the record indicates that the target popu- lation is responding rationally. DHS’s system of counting and stacking benefits is hardly transparent, and so a rational per- son might err on the side of caution and refrain from seeking medical care, or food, or housing, even from a city, state, or tribe rather than the federal government. And the risk that the Rule may become more stringent at any time and operate ret- roactively against the use of benefits already used is a real one. DHS trumpets its view that the Rule stops short of its lawful authority and that it could promulgate a more restric- tive rule if it so chooses. In response to comments on the pro- posed rule, DHS used discretionary language: “DHS believes it is a reasonable approach to only designate Medicaid at this time,” id. at 41381 (emphasis added); and “DHS will not con- sider [Healthy Start] benefits at this time,” id. at 41390 (empha- sis added). It warned that it may “updat[e] the list of benefits through future regulatory action.” Id. at 41387. Immigrants thus reasonably anticipate that their receipt of benefits that are currently not covered could eventually hurt them if DHS alters the Rule in the future. It was not enough for DHS simply to nod at this argument; it called for a serious explanation. The importance of the chilling effect is not the number of disenrollments in the ab- stract, but the collateral consequences of such disenrollments. DHS failed adequately to grapple with the latter. For example, commenters predicted that disenrollment and under-enroll- ment in Medicaid, including by immigrants not covered by the Rule, would reduce access to vaccines and other medical care, resulting in an increased risk of an outbreak of infectious disease among the general public. To recognize the truth in No. 19-3169 35 that prediction, one need only consider the current outbreak of COVID-19—a pandemic that does not respect the differ- ences between citizens and noncitizens. There is also the added burden on states and local govern- ments, which must disentangle their purely state-funded pro- grams from covered federal programs. The federal govern- ment has no interest in the way that states and localities choose to spend their money. There is no reason why immi- grants should not continue to benefit from the state programs without being penalized at the federal level. The Rule will force states to make their own public welfare programs more robust to compensate for a reduction in the availability of fed- eral programs. DHS touts the savings to the federal govern- ment from the Rule, primarily through a significant reduction in transfer payments to the states (including, it should be noted, for persons who disenroll unnecessarily because of the chilling effect), but at the same time it expects the states to fill the gaps and continue to provide critical services such as pre- ventive healthcare. See, e.g., id. at 41385 (“In addition, local health centers and state health departments provide preven- tive services that include vaccines that may be offered on a sliding scale fee based on income. Therefore, DHS believes that vaccines would still be available for children and adults even if they disenroll from Medicaid.”). It assumes this while simultaneously denying that the Rule will have “substantial direct effects on the States, on the relationship between the Federal Government and the States, or on the distribution of power and responsibilities among the various levels of gov- ernment.” Id. at 41481. Cook County also asserts that DHS failed to give a logical rationale for its chosen durational threshold. In its notice of 36 No. 19-3169 proposed rulemaking, DHS proposed an array of thresholds that would apply before benefits can be counted against a noncitizen in the public charge analysis. Those lines came un- der sharp criticism for being arbitrary, confusing, and an un- acceptable proxy for undue reliance on public support. Id. at 41357–58. In the final Rule, DHS opted for the single threshold for both monetizable and nonmonetizable benefits of 12 months (stacked) over a 36-month period. It touted this approach as “particularly responsive to public comments that communi- cated concerns about the complexity of the bifurcated stand- ard and lack of certainty.” Id. at 41358. It also asserted that the 12/36 standard “is consistent with DHS’s interpretation of the term ’public charge.’” Id. at 41359. DHS equates the term “public charge” with a lack of “self-sufficiency” and it regards anyone who fails its test as not self-sufficient. Id. It defends its stacking mandate on the theory that it “ensures that aliens who receive more than one public benefit (which may be more indicative of a lack of self-sufficiency, with respect to the ful- fillment of multiple types of basic needs) reach the 12-month limit faster.” Id. at 41361. DHS concluded that the bright-line rule “provides meaningful guidance to aliens and adjudica- tors, … accommodates meaningful short-term and intermit- tent access to public benefits, and … does not excuse continu- ous or consistent public benefit receipt that denotes a lack of self-sufficiency.” Id. This explains how DHS incorporated its understanding of “self-sufficiency” into the Rule. But we still have a textual problem. The INA does not call for total self-sufficiency at every moment; it uses the words “public charge.” DHS sees “lack of complete self-sufficiency” and “public charge” as No. 19-3169 37 synonyms: in its view, receipt of any public benefit, particu- larly one related to core needs such as health care, housing, and nutrition, shows that a person is not self-sufficient. See id. at 41356. This is an absolutist sense of self-sufficiency that no person in a modern society could satisfy; everyone relies on nonmonetary governmental programs, such as food safety, police protection, and emergency services. DHS does not offer any justification for its extreme view, which has no basis in the text or history of the INA. As we explained earlier, since the first federal immigration law in 1882, Congress has as- sumed that immigrants (like others) might face economic in- security at some point. Instead of penalizing immigrants by denying them entry or the right to adjust status, Congress built into the law accommodations for that reality. Also, as numerous commenters on the Rule pointed out, the benefits it covers are largely supplemental and not intended to be, or relied upon as, a primary resource for recipients. Many recip- ients could get by without them, though as a result they would face greater health, nutrition, and housing insecurity, which in turn would likely harm their work or educational attainment (and hence their ability to be self-sufficient). Finally, Cook County contends that the Rule adds irra- tional factors into the public-charge assessment, including family size, mere application for benefits, English-language proficiency, lack of disability, and good credit history. With respect to language, we note the obvious: someone whose English is limited on the date of entry may be entirely compe- tent five years later, when the person first becomes eligible for benefits under the Welfare Reform Act and related laws. In almost all cases, an immigration official making a determina- tion about whether someone is likely to become a public 38 No. 19-3169 charge will be speculating about that person’s family size, lin- guistic abilities, credit score, and the like no fewer than five years in the future. Even if we grant that these new factors carry some mini- mal probative value, it is unclear to us, and DHS nowhere ex- plains, how immigration officials are supposed to make these predictions in a nonarbitrary way. Worse, for many people the relevant time is not five years—it is eternity, because the Rule calls for officials to guess whether an immigrant will be- come a public charge at any time. There is a great risk that of- ficials will make their determination based on stereotype or unsupported assumptions, rather than on the type of objec- tive facts called for by the Act (age, present health, family sta- tus, financial situation, and education or skills). DHS also never explains why it chose not to take into ac- count the possibility that an immigrant might, at some point in the future, be able to repay the value of public benefits re- ceived. Someone who seeks to adjust status will be penalized for having previously received public benefits without being given the opportunity to refund the government the cost of those benefits. This is new: the regulations governing depor- tation on public-charge grounds require a demand and a fail- ure to pay. See 64 Fed. Reg. at 28691. All of this convinces us that this Rule is likely to fail the “arbitrary and capricious” standard. The Rule has numerous unexplained serious flaws: DHS did not adequately consider the reliance interests of state and local governments; did not acknowledge or address the significant, predictable collateral consequences of the Rule; incorporated into the term “public charge” an understanding of self-sufficiency that has no basis in the statute it supposedly interprets; and failed to address No. 19-3169 39 critical issues such as the relevance of the five-year waiting period for immigrant eligibility for most federal benefits. B. Other Criteria for Preliminary Injunction We have spent most of our time on likelihood of success on the merits, because that is the critical factor here. We add only a few words about the other requirements for prelimi- nary relief. Cook County had to show that it is likely to suffer irreparable harm in the absence of preliminary relief; that le- gal remedies are inadequate; and that the balance of equities tips in their favor. The district court found that it did so. As we noted earlier, Cook County has shown that the Rule will cause immigrants, including those not covered by the Rule, to disenroll from, or refrain from enrolling in, federal Medicaid and state-level public health programs. This already has led to reduction in rates of preventive medicine and caused immigrants to rely on uncompensated emergency care from Cook County’s hospital system; the record supports the prediction that those harms will only get worse. The result for the County will be a significant increase in costs it must bear and a higher county-wide risk of vaccine-preventable and other communicable diseases for its population as a whole. The record also supports the district court’s finding that Cook County will have to divert resources away from existing pro- grams to respond to the effects of the Rule. The district court was also on solid ground in finding that Cook County lacks adequate legal remedies for the injuries imposed by the Rule. The APA provides a limited waiver of the United States’ sovereign immunity and supports a claim for a challenge to agency action, but only to the extent that the plaintiffs “seek relief other than money damages.” 5 U.S.C. 40 No. 19-3169 § 702. There is thus no post-hoc legal remedy available to Cook County to redress the financial harms it stands to suffer as a result of the Rule. It is injunctive relief or nothing. With respect to the balance of harms, we must take ac- count of the Supreme Court’s decision to stay the preliminary injunction entered by the district court. The Court’s stay deci- sion was not a merits ruling. To succeed in obtaining a stay from the Supreme Court, an applicant “must demonstrate (1) a reasonable probability that four Justices will consider the is- sue sufficiently meritorious to grant certiorari or to note prob- able jurisdiction; (2) a fair prospect that a majority of the Court will conclude that the decision below was erroneous; and (3) a likelihood that irreparable harm will result from the denial of a stay.” Conkright v. Frommert, 556 U.S. 1401, 1402 (2009) (Ginsburg, J., in chambers). Stays, the Court tells us, are “granted only in extraordinary cases.” Id. We do not know why the Court granted this stay, because it did so by sum- mary order, but we assume that it abided by the normal standards. Consequently, the stay provides an indication that the Court thinks that there is at least a fair prospect that DHS should prevail and faces a greater threat of irreparable harm than the plaintiffs. The stay thus preserves the status quo while this case and others percolate up from courts around the country. There would be no point in the merits stage if an issuance of a stay must be understood as a sub silentio disposition of the under- lying dispute. With the benefit of more time for consideration and the complete preliminary injunction record, we believe that it is our duty to evaluate each of the preliminary injunc- tion factors, including the balance of equities. In so doing, we apply a “sliding scale” approach in which “the more likely the No. 19-3169 41 plaintiff is to win, the less heavily need the balance of harms weigh in his favor; the less likely he is to win, the more need it weigh in his favor.” Valencia v. City of Springfield, 883 F.3d at 966. We also consider effects that granting or denying the pre- liminary injunction would have on the public. Id. In our view, Cook County has shown that it is likely to suffer (and has already begun to suffer) irreparable harm caused by the Rule. Given the dramatic shift in policy the Rule reflects and the potentially dire public health consequences of the Rule, we agree with the district court that the public inter- est is better served for the time being by preliminarily enjoin- ing the Rule. IV. Conclusion While we disagree with the district court that this case can be resolved at step one of the Chevron analysis, we agree that at least Cook County has standing to sue. We make no ruling on ICIRR’s standing, and so we have based the remainder of our opinion on Cook County’s situation only. The district court did not abuse its discretion or err as a matter of law when it concluded that Cook County is likely to succeed on the merits of its APA claims against DHS. Nor did the district court’s handling of the balance of harms and lack of alterna- tive legal remedies represent an abuse of discretion. We there- fore AFFIRM the district court’s order entering a preliminary injunction. 42 No. 19-3169 BARRETT, Circuit Judge, dissenting. The plaintiffs have worked hard to show that the statutory term “public charge” is a very narrow one, excluding only those green card applicants likely to be primarily and perma- nently dependent on public assistance. That argument is be- lied by the term’s historical meaning—but even more im- portantly, it is belied by the text of the current statute, which was amended in 1996 to increase the bite of the public charge determination. When the use of “public charge” in the Immi- gration and Nationality Act (INA) is viewed in the context of these amendments, it becomes very difficult to maintain that the definition adopted by the Department of Homeland Secu- rity (DHS) is unreasonable. Recognizing this, the plaintiffs try to cast the 1996 amendments as irrelevant to the meaning of “public charge.” That argument, however, flies in the face of the statute—which means that despite their best efforts, the plaintiffs’ interpretive challenge is an uphill battle that they are unlikely to win. I therefore disagree with the majority’s conclusion that the plaintiffs’ challenge to DHS’s definition of “public charge” is likely to succeed at Chevron step two. I express no view, how- ever, on the majority’s analysis of the plaintiffs’ other chal- lenges to the rule under the Administrative Procedure Act. The district court did not reach them, and the plaintiffs barely briefed them. The preliminary injunction was based solely on the district court’s interpretation of the term “public charge.” Because its analysis was flawed, I would vacate the injunction and remand the case to the district court, where the plaintiffs would be free to develop their other arguments. No 19-3169 43 I. There is a lot of confusion surrounding the public charge rule, so I’ll start by addressing who it affects and how it works. The plaintiffs emphasize that the rule will prompt many noncitizens to drop or forgo public assistance, lest their use of benefits jeopardize their immigration status. That’s happening already, and it’s why Cook County has standing: noncitizens who give up government-funded healthcare are likely to rely on the county-funded emergency room. But it’s important to recognize that immigrants are dropping or for- going aid out of misunderstanding or fear because, with very rare exceptions, those entitled to receive public benefits will never be subject to the public charge rule. Contrary to popular perception, the force of the rule does not fall on immigrants who have received benefits in the past. Rather, it falls on nonimmigrant visa holders who, if granted a green card, would become eligible for benefits in the future. To see why, one must be clear-eyed about the fact that fed- eral law is not particularly generous about extending public assistance to noncitizens. That is not a function of the public charge rule; it is a function of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996, Pub. L. No. 104-193, 110 Stat. 2105 (1996), commonly referred to as the “Welfare Reform Act.” Under the Act, undocumented noncit- izens are ineligible for benefits. So are nonimmigrant visa holders, a category that encompasses noncitizens granted permission to be in the United States for a defined period— think of tourists, students, and temporary workers. See 8 U.S.C. §§ 1611(a), 1621(a), 1641(b) (excluding undocumented noncitizens and nonimmigrant visa holders from the list of 44 No. 19-3169 noncitizens “qualified” for government benefits). 1 Because of these restrictions, many noncitizens are altogether ineligible for the benefits relevant to a public charge determination. Only two major groups are statutorily eligible to receive the benefits that the public charge rule addresses, and the rule has little to no effect on either. The first group is certain espe- cially vulnerable populations—refugees and asylees, among others. Congress has entitled these vulnerable noncitizens to public assistance, 8 U.S.C. § 1641(b), and exempted them from the public charge exclusion, id. §§ 1157(c)(3), 1159(c). That means that their need for aid is not considered when they are admitted to the United States, nor is their actual receipt of aid considered in any later adjustment-of-status proceeding. The 1 There are some narrow exceptions, but they are irrelevant to the “public charge” determination. All noncitizens, including the undocu- mented, are eligible to receive short-term, in-kind emergency disaster re- lief; certain forms of emergency medical assistance; public-health assis- tance for immunization, as well as treatment for the symptoms of com- municable disease; other in-kind services such as soup kitchens and crisis counseling; and housing benefits to the extent that the noncitizen was re- ceiving public housing prior to 1996. 8 U.S.C. §§ 1611(b), 1621(b). Other than the housing benefits, none of this aid counts under the rule’s defini- tion of a “public benefit,” so none has any effect on any future adjustment- of-status proceeding. See 8 C.F.R. § 212.21; see also Inadmissibility on Pub- lic Charge Grounds, 84 Fed. Reg. 41,292, 41,313 (Aug. 14, 2019) (noting that the rule’s “definition does not include benefits related exclusively to emer- gency response, immunization, education, or social services”); id. at 41,482 (explaining that the rule’s definition “does not include emergency aid, emergency medical assistance, or disaster relief”). And while housing benefits are covered by the public charge rule, 8 C.F.R. § 212.21, they are largely irrelevant because the number of noncitizens still within the grandfathering provision has presumably dwindled dramatically in the quarter century since the Welfare Reform Act was passed. No 19-3169 45 public charge rule is entirely irrelevant to the most vulnera- ble. The second group eligible for benefits is lawful permanent residents, often referred to as green card holders, and the rule is almost entirely irrelevant to them too. Here’s why: The pub- lic charge exclusion applies to noncitizens at the admission stage or an adjustment-of-status proceeding. Id. § 1182(a)(4)(A). (“Admission” is a term of art referring to “the lawful entry of the alien into the United States after inspection and authorization by an immigration officer.” Id. § 1101(a)(13)(A).) Lawful permanent residents have already been admitted to the United States, and they already possess the most protected immigrant status. They are therefore not subject to the public charge exclusion unless they jeopardize their lawful permanent residency. See id. § 1101(a)(13)(C) (de- scribing the narrow circumstances in which lawful perma- nent residents are considered to be “seeking an admission”). Most relevant here, a green card holder who leaves the coun- try for more than 180 days puts her residency in question and might need to “seek[] an admission” upon returning to the United States. Id. § 1011(a)(13)(C)(iii). If she used benefits prior to her departure, then her use of those benefits might count against her at reentry. But this consequence is easy to avoid by keeping trips abroad shorter than six months. It’s also worth noting that a lawful permanent resident is eligible to receive very few benefits until she has been here for five years—which is the point at which she is eligible for citizen- ship. Id. § 1427(a). Naturalization eliminates even the small risk that a lawful permanent resident would ever face the ad- mission process again. Notably, the rule doesn’t apply at the naturalization stage. See id. § 1429. 46 No. 19-3169 The upshot is that the public charge rule will rarely apply to a noncitizen who has received benefits in the past. 2 Indeed, in the Second Circuit case challenging this same rule, both the government and the plaintiffs conceded as much. When pressed to identify who could be penalized under the public charge rule for using benefits, neither side identified any ex- ample other than the 180-day departure of a lawful perma- nent resident. See Oral Argument at 36:06–38:47, 1:03:45– 1:04:40, New York v. U.S. Dep’t of Homeland Sec., Nos. 19-3591, 19-3595 (2d Cir. Mar. 2, 2020), https://www.c- span.org/video/?469804-1/oral-argument-trump-administra- tion-public-charge. Notwithstanding all of this, many lawful permanent resi- dents, refugees, asylees, and even naturalized citizens have disenrolled from government-benefit programs since the public charge rule was announced. Given the complexity of immigration law, it is unsurprising that many are confused or fearful about how the rule might apply to them. Still, the pat- tern of disenrollment does not reflect the rule’s actual scope. Focusing on the source of Cook County’s injury can therefore be misleading. That does not mean, however, that the rule has no effect. Even though it is almost entirely inapplicable to those cur- rently eligible for benefits, it significantly affects a different group: nonimmigrant visa holders applying for green cards. 2Hence the majority is wrong to treat the rule as unreasonable be- cause it “set[s] a trap for the unwary.” Maj. Op. at 29. Because those eligi- ble for the designated benefits are not subject to the rule—except in very rare circumstances—it does not “penaliz[e] people for accepting benefits Congress made available to them.” Id. No 19-3169 47 Recall that nonimmigrant visa holders, unlike lawful perma- nent residents and those holding humanitarian-based visas, are ineligible for the relevant benefits in their current immi- gration status. If granted lawful permanent residency, though, they would become eligible for these benefits in the future. The public charge rule is concerned with what use a green card applicant would make of this future eligibility. As a leading treatise puts it, the public charge determination is a “prophetic” one. 5 CHARLES GORDON ET AL., IMMIGRATION LAW AND PROCEDURE § 63.05[3] (2019). If DHS predicts that an applicant is likely to rely too heavily on government assis- tance, it will deny her lawful permanent residency on the ground that she is likely to become a public charge. This case is about whether DHS has defined “public charge” too expan- sively and is therefore turning too many noncitizens away. There are four major routes to obtaining the status of law- ful permanent resident: humanitarian protection (refugees and asylees), the sponsorship of a family member, employ- ment, and winning what is known as the green card lottery. 3 See U.S. DEP’T OF HOMELAND SEC., OFFICE OF IMMIGRATION STATISTICS, ANNUAL FLOW REPORT: LAWFUL PERMANENT RESIDENTS 3–4 (2018), https://www.dhs.gov/sites/de- fault/files/publications/Lawful_Permanent_Resi- dents_2017.pdf. Those seeking humanitarian protection are not subject to the statutory provision rendering inadmissible any “alien who … is likely at any time to become a public charge,” 8 U.S.C. § 1182(a)(4)(A), and only a subset of those in the remaining three categories will be subject to the DHS rule. 3 The diversity visa, commonly referred to as the green card lottery, is awarded to foreign nationals from underrepresented countries in an effort to increase diversity within the United States. See 8 U.S.C. § 1153(c). 48 No. 19-3169 That is because DHS only handles the applications of noncit- izens who apply from within the United States; the State De- partment processes the applications of noncitizens who apply from abroad. 4 This division of authority means that, as a prac- tical matter, the regulation applies to those present in the United States on nonimmigrant visas who seek to adjust their status to that of lawful permanent residents. And because the green card lottery is processed almost entirely by the State De- partment, the DHS rule applies primarily to employment- based applicants and family-based applicants (by far the larger of these two groups). 5 As nonimmigrant visa holders, these applicants have not previously been eligible for the benefits designated by DHS’s rule—so the determination is not a backward-looking inquiry into whether they have used such benefits in the past. Instead, it is a forward-looking inquiry into whether they are likely to use such benefits in the future. The rule guides this forward- looking inquiry. Under the 1999 Guidance, an applicant was 4 The State Department has adopted the interpretation set forth in this rule, but its implementation of the public charge exclusion is not at issue in this case. See Visas: Ineligibility on Public Charge Grounds, 84 Fed. Reg. 54,996, 55,000 (Oct. 11, 2019). 5 In 2019, approximately 572,000 noncitizens adjusted their status to that of lawful permanent residents. The largest group—roughly 330,000— were family based, and the majority of those (over 217,000) were spouses of U.S. citizens. About 111,000 were employment based, and only about 1,000 were lottery winners. The vast majority of the remaining 130,000 noncitizens—refugees and asylees, among others—were exempt from the public charge rule. See Legal Immigration and Adjustment of Status Report Data Tables: FY 2019, U.S. DEP’T HOMELAND SECURITY tbl.1B (Jan. 15, 2020), https://www.dhs.gov/immigration-statistics/readingroom/spe- cial/LIASR#. No 19-3169 49 excluded only if she was likely to be institutionalized or pri- marily dependent on government cash assistance for the long term. Field Guidance on Deportability and Inadmissibility on Public Charge Grounds, 64 Fed. Reg. 28,689, 28,689 (Mar. 26, 1999). Now, DHS considers the applicant’s potential usage not only of cash assistance for income maintenance (including Temporary Assistance for Needy Families (TANF), Supple- mental Security Income (SSI), and state cash assistance), but also of the Supplemental Nutrition Assistance Program (SNAP), the Section 8 Housing Choice Voucher Program, Sec- tion 8 project-based rental assistance, housing benefits under Section 9, and Medicaid (with some explicit exceptions). 8 C.F.R. § 212.21. And if DHS concludes that an applicant is likely to use more than 12 months’ worth of these benefits— with the use of 2 benefits in 1 month counting as 2 months— it will deem her “likely to become a public charge” and deny the green card. Id. This heightened standard for admissibility is a significant change—but it’s not the one that the plaintiffs’ emphasis on disenrollment suggests. Evaluating the rule requires a clear view of what it actually does; so, with the rule’s scope in mind, I turn to the merits. II. While I agree with the majority’s bottom-line conclusion at Chevron step one that “public charge” does not refer exclu- sively to one who is primarily and permanently dependent on government assistance, I have a little to add to the history and a lot to add to the statutory analysis. In my view, the majority takes several wrong turns in analyzing the statute that skew its thinking about Chevron step two. For purposes of this Part, the most significant is that the majority accepts the plaintiffs’ 50 No. 19-3169 view that the 1996 amendments to the public charge provision were irrelevant. In what follows, I’ll lay out my own analysis of the plaintiffs’ arguments, which will explain why I wind up in a different place than the majority does on the reasona- bleness of DHS’s interpretation of the statute. The plaintiffs advance three basic arguments as to why the term “public charge” refers exclusively to one who is “pri- marily and permanently” dependent on government assis- tance. First, they say that the term had that meaning when it first appeared in the 1882 federal statute. Second, they con- tend that even if the term was unsettled in the late nineteenth century, subsequent judicial and administrative decisions narrowed it, and later amendments to the statute ratified these interpretations. Third, they argue that interpreting the term “public charge” to encompass anything short of primary and permanent dependence conflicts with Congress’s choice to make supplemental government benefits available to im- migrants. I’ll take these arguments in turn. A. The plaintiffs first argue that in the late nineteenth cen- tury, “public charge” meant primary and permanent depend- ence. See New Prime Inc. v. Oliveira, 139 S. Ct. 532, 539 (2019) (“[I]t’s a ‘fundamental canon of statutory construction’ that words generally should be ‘interpreted as taking their ordi- nary … meaning … at the time Congress enacted the stat- ute.’” (citation omitted)). Evaluating this argument requires careful consideration of a term with a long history. The term “public charge” was borrowed from state “poor laws,” which were in turn modeled on their English counterparts. HIDETAKA HIROTA, EXPELLING THE POOR: ATLANTIC SEABOARD STATES AND THE NINETEENTH-CENTURY ORIGINS OF AMERICAN No 19-3169 51 IMMIGRATION POLICY 43–47 (2017). Early poor laws used “public charge” synonymously with “public expense,” refer- ring to any burden on the public fisc. Thus, when someone sought assistance from a city or county overseer of the poor, the cost of the relief provided was entered on the overseer’s books as a public charge—that is, an expense properly charge- able to, and therefore funded by, the public. Over time, the term “public charge” came to refer (at least in the context of poor relief and immigration laws) not only to expenditures made under the poor laws, but also to the people who de- pended on these expenditures. 6 State legislatures, worried about the burden that destitute immigrants might place on programs to aid the needy, co- opted the poor-law language into immigration legislation. In 1847, New York created an administrative apparatus for deal- ing with the influx of immigrants. The new “Commissioners of Emigration” were tasked with examining incoming passen- gers to determine if “there shall be found among such passen- gers, any lunatic, idiot, deaf and dumb, blind or infirm per- sons … who, from attending circumstances, are likely to be- come permanently a public charge”—language, incidentally, that suggests that one could be a public charge either tempo- rarily or permanently. Act of May 5, 1847, ch. 195, § 3, 1847 N.Y. Laws 182, 184. These individuals were permitted to land in the state upon payment of a bond by the vessel’s master “to indemnify … each and every city, town and county within 6 This is why nineteenth-century dictionary definitions of “charge” are unhelpful. The words “public” and “charge” comprise a unit that must be understood in the context of the laws that used the phrase. Cf. Yates v. United States, 135 S. Ct. 1074, 1082 (2015) (“[A]lthough dictionary defini- tions of the words ‘tangible’ and ‘object’ bear consideration, they are not dispositive of the meaning of ‘tangible object’ ….”). 52 No. 19-3169 this state, from any cost or charge … for the maintenance or support of the person … within five years.” Id. The bonds paid for the landing of these immigrants were then used to pay for the state immigration infrastructure, including the provision of some temporary aid to new arrivals. Two years later, the state expanded the category of people for whom a bond was required. Still excluded were those “likely to be- come permanently a public charge” but also those “who have been paupers in any other country, or who from sickness or disease, existing at the time of departing from the foreign port, are or are likely to soon become a public charge.” Act of Apr. 11, 1849, ch. 350, § 3, 1849 N.Y. Laws 504, 506. By 1851, the New York statute contained the language which would be included in both the 1882 and 1891 federal statutes. Gone was the reference to those “likely to become permanently a public charge,” replaced by phrases referring to someone “unable to take care of himself or herself without becoming a public charge” and someone “likely to become a public charge.” Act of July 11, 1851, ch. 523, § 4, 1851 N.Y. Laws 969, 971. In the event that a bond was unpaid, New York—and Massachu- setts, which enacted a substantially similar law—ordered the exclusion of those immigrants deemed “likely to soon become a public charge.” HIROTA, supra, at 71–72. The bond system was held unconstitutional by the U.S. Su- preme Court on the ground that that the power to tax incom- ing foreign passengers “has been confided to Congress by the Constitution.” Henderson v. Mayor of New York, 92 U.S. 259, 274 (1876). The decision threw the state systems into uncertainty and created demand for federal legislation, largely to reenact the defunct state policies and to replace the lost funding. Since the states could no longer fund their immigration systems us- ing state bonds, the 1882 federal statute levied “a duty of fifty No 19-3169 53 cents for each and every passenger not a citizen of the United States” arriving by sea; this was to “constitute a fund … to defray the expense of regulating immigration … and for the care of immigrants arriving in the United States, for the relief of such as are in distress.” Act of Aug. 3, 1882, ch. 376, § 1, 22 Stat. 214, 214. The first federal statute therefore filled the space left by the now-ineffective state laws: it used funds raised from the immigrants or their carriers to provide some care for the newly arrived, while describing criteria for excluding those likely to financially burden state and local governments. Because the term “public charge” had been pulled directly from the state statutes, it presumably had the same meaning that it had come to have under the state laws: someone who depended, or would likely depend, on poor-relief programs. But when the term “public charge” was imported into fed- eral law, it was unclear how much state aid qualified someone as a “public charge.” Neither state poor laws nor state immi- gration laws defined “public charge,” and no clear definition emerged in judicial opinions or secondary sources, either. Early efforts to enforce the 1882 statute bear out the uncer- tainty surrounding the term. In 1884, an association of ten steamship companies asked the Secretary of the Treasury, on whom responsibility for immigration fell at the time, to “spe- cifically define … the circumstances which shall constitute ‘a person unable to take care of himself or herself without be- coming a public charge,’ and who shall not be permitted to land under … the [1882] act.” SYNOPSIS OF THE DECISIONS OF THE TREASURY DEPARTMENT ON THE CONSTRUCTION OF THE TARIFF, NAVIGATION, AND OTHER LAWS FOR THE YEAR ENDED DECEMBER 31, 1884, at 365 (1885). (The steamship companies had a stake because they were on the hook for the noncitizen’s return ticket if she was rejected as a likely public charge.) The 54 No. 19-3169 Secretary demurred, answering that “the determination of the liability of arriving immigrants to become public charges is vested … in the commissioners of immigration appointed by the State in which such immigrants arrive,” and thus “this De- partment must decline to interfere in the matter.” Id. One year later, Treasury continued to recognize that “difficulties have arisen in regard to the construction of so much of section 2 of [the 1882 act] … as refers to the landing of convicts, lunatics, idiots, or persons unable to take care of themselves without becoming a public charge,” though it still refused to offer clar- ification. SYNOPSIS OF THE DECISIONS OF THE TREASURY DEPARTMENT ON THE CONSTRUCTION OF THE TARIFF, NAVIGATION, AND OTHER LAWS FOR THE YEAR ENDING DECEMBER 31, 1885, at 359 (1886). The term was not necessarily clarified in 1891, when im- migration-enforcement authority was placed directly in the hands of federal officials. (From 1882 until Congress enacted the Immigration Act of 1891, states had continued to admin- ister immigration enforcement, albeit under authority con- ferred by the federal statute.) With the change in administra- tion, the steamship companies continued to express confu- sion, informing Treasury officials that the phrase “was some- what indefinite and [that they] desired to have a more specific explanation of its meaning.” 1 LETTER FROM THE SECRETARY OF THE TREASURY, TRANSMITTING A REPORT OF THE COMMISSIONERS OF IMMIGRATION UPON THE CAUSES WHICH INCITE IMMIGRATION TO THE UNITED STATES 109 (1892). At this point, Treasury offered an answer, but it was hardly clarify- ing. Pressed by Congress to describe the standards used by officials to determine whether an immigrant was “likely to be- come a public charge,” the Assistant Secretary in 1892 re- sponded that “written instructions and an inflexible standard No 19-3169 55 would be inapplicable and impracticable … and the sound discretion of the inspection officer, subject to appeal as pre- scribed by law, must be the chief reliance.” H.R. REP. NO. 52- 2090, at 4 (1892). Rather than conveying something narrow and definite, the term “public charge” seemed to refer in an imprecise way to someone who lacked self-sufficiency and therefore burdened taxpayers. Explanations of the term offered in a congressional hearing by John Weber, the first commissioner of immigration at Ellis Island, illustrate the point. He explained that “[t]he appearance of the man, his vocation, his willingness to work, his apparent industry, and the demand for the kind of work that he is ready to give, is what governs” whether an individ- ual was likely to become a public charge. Id. at 359. When asked whether an immigrant would be considered likely to become a public charge if “it is necessary that a private charity shall furnish food and lodging … for a period long or short after landing,” Weber responded that such a person would likely be considered a public charge, but that it would not vi- olate the statute to allow him to land so long as it was obvious that he would be “supported on private charity only up to the time when [he got] employment, which may only be until the next day.” Id. at 425. The repeated requests for clarification from steamship op- erators and Congress, coupled with Treasury’s reluctance to provide a concrete answer, indicate that the term did not have a definite and fixed meaning. That is unsurprising in the con- text of the time: it would have been difficult to have a one- size-fits-all definition of how much aid was too much, be- cause there was not a one-size-fits-all system of welfare. Poor relief was largely handled by towns and counties, which 56 No. 19-3169 made their own choices about how to deliver aid. Most local- ities deployed “outdoor relief”—in-kind and cash support without institutionalization. See MICHAEL B. KATZ, IN THE SHADOW OF THE POORHOUSE: A SOCIAL HISTORY OF WELFARE IN AMERICA 37 (1986) (“[P]oorhouses did not end public outdoor relief. With a few exceptions, most towns, cities, and counties helped more people outside of poorhouses than within them.”). Other areas were more reliant on “indoor” relief in the form of poorhouses. Id. at 16–18. Some used a mixed sys- tem, adjusting the provision of indoor and outdoor relief as poorhouse populations ebbed and flowed. Id. at 39. And while the plaintiffs treat residence in a poorhouse as a proxy for primary and permanent dependence, that’s not how poor- houses worked—they housed a mix of the permanently and temporarily dependent, serving as “both a short-term refuge for people in trouble and a home for the helpless and elderly.” Id. at 90. The bottom line is that in the closing decades of the nine- teenth century, several different forms of public relief existed contiguously. And when nineteenth-century immigration of- ficials determined whether someone was “likely to become a public charge,” dependence on a particular kind or amount of relief does not appear to have been dispositive. Rather than serving as shorthand for a certain type or duration of aid, the term “public charge” referred to a lack of self-sufficiency that officials had broad discretion to estimate. Neither state legis- latures nor Congress pinned down the term any more than that. B. The plaintiffs have a backup argument: even if the term was unsettled in the late nineteenth century, they claim that it No 19-3169 57 became settled in the twentieth. According to the plaintiffs, courts and administrative agencies repeatedly held that “pub- lic charge” meant one who is “primarily and permanently de- pendent” on the government, and Congress ratified this set- tled meaning in its many reenactments of the public charge provision. See WILLIAM N. ESKRIDGE JR., INTERPRETING LAW: A PRIMER ON HOW TO READ STATUTES AND THE CONSTITUTION app. at 421 (2016) (“When Congress reenacts a statute, it in- corporates settled interpretations of the reenacted statute.”). Thus, the plaintiffs say, whatever uncertainty may have sur- rounded the term in 1882, there was no uncertainty when Congress reenacted the provision. And because Congress reenacted the provision many times—in 1891, 1907, 1917, 1952, 1990, and 1996—the plaintiffs canvass a century’s worth of judicial and administrative precedent in an effort to show that a consensus existed before at least one of these reenact- ments. The bar for establishing a settled interpretation is high: at the time of reenactment, the judicial consensus must have been “so broad and unquestioned that we must presume Con- gress knew of and endorsed it.” Jama v. Immigration & Customs Enf’t, 543 U.S. 335, 349 (2005). The plaintiffs rely heavily on Gegiow v. Uhl, 239 U.S. 3 (1915), to establish this consensus, but I share the majority’s view that Gegiow doesn’t do the work that the plaintiffs want it to. In that case, the Court did not define “public charge” other than to say that it cannot be defined with reference to labor conditions in the city in which an immigrant intends to settle. The Court concluded that im- migrant arrivals “are to be excluded on the ground of perma- nent personal objections accompanying them irrespective of local conditions unless the one phrase before us [public charge] is directed to different considerations than any other 58 No. 19-3169 of those with which it is associated.” Id. at 10. In other words, classifying someone as a likely “public charge” does not de- pend on whether he is bound for Portland or St. Paul. The Court did not define the degree of reliance that renders some- one a “public charge,” because that was not the question be- fore it. Thus, Gegiow neither binds us nor offers a definition that Congress could have ratified. 7 Without Gegiow, the plaintiffs face an uphill battle because satisfying the requirements of the reenactment canon typi- cally requires at least one Supreme Court decision. See, e.g., Forest Grove Sch. Dist. v. T.A., 557 U.S. 230, 244 n.11 (2009); Cannon v. Univ. of Chi., 441 U.S. 677, 699 (1979). And for the reasons that the majority gives, this is not the rare case in which lower court and administrative decisions are enough to demonstrate a consensus. See Maj. Op. at 21–22; see also Wil- liam N. Eskridge, Jr., Interpreting Legislative Inaction, 87 MICH. L. REV. 67, 83 (1988) (“[T]he Court often will not incorporate lower court decisions into a statute through the reenactment rule.”); cf. Tex. Dep’t of Hous. & Cmty. Affairs v. Inclusive Cmtys. 7 It is worth noting that even after Gegiow, state and local governments took varied positions on what it meant for an immigrant to be a public charge. For instance, in the 1920s, Los Angeles worked closely with chari- table institutions to report as public charges immigrants who were receiv- ing outdoor relief. Cybelle Fox, The Boundaries of Social Citizenship: Race, Immigration and the American Welfare State, 1900–1950, at 266–67 (May 7, 2007) (unpublished Ph.D. dissertation, Harvard University). But other jurisdictions rarely reported immigrants who were receiving only outdoor relief—for example, as early as the 1920s, Cook County devel- oped its own local policy to not “deport when the necessity for public care [was] only temporary.” Id. at 278. No 19-3169 59 Project, Inc., 135 S. Ct. 2507, 2520 (2015) (applying the reenact- ment canon in light of “the unanimous holdings of the Courts of Appeals”) (emphasis added). In any event, the reenactment canon requires more than a judicial consensus—it applies only if Congress reenacted the provision without making material changes. Jama, 543 U.S. at 349; see also Holder v. Martinez Gutierrez, 566 U.S. 583, 593 (2012) (“[T]he doctrine of congressional ratification applies only when Congress reenacts a statute without relevant change.”). Whatever one thinks of earlier changes to the pub- lic charge provision, there can be no doubt that the 1996 amendments were material. The INA is notoriously complex, and these amendments are no exception. Making matters worse, the amendments came from two separate acts, themselves incredibly complex, that were passed a month apart: the Welfare Reform Act, Pub. L. No. 104-193, 110 Stat. 2105 (1996), and the Illegal Immigra- tion Reform and Immigrant Responsibility Act, Pub. L. No. 104-208, 110 Stat. 3009-546 (1996) (IIRIRA). But because the plaintiffs challenge the materiality of these amendments to the meaning of the term “public charge,” it is necessary to step through them at a level of detail that is, unfortunately, excru- ciating. Congress enacted IIRIRA, which made sweeping changes to the INA, in September of 1996. Among its changes were several material amendments to the public charge provision. For the first time in the provision’s 114-year history, Congress required the Executive to consider an itemized list of factors in making the public charge determination, thereby ensuring that the inquiry was searching rather than superficial. See 8 U.S.C. § 1182(a)(4)(B)(i) (providing that “the consular officer 60 No. 19-3169 or the Attorney General shall at a minimum consider” the noncitizen’s age; health; family status; assets, resources, and financial status; and education and skills). Even more signifi- cantly, it added a subsection to the public charge provision rendering most family-sponsored applicants automatically inadmissible on public charge grounds unless they obtained an enforceable affidavit of support from a sponsor (usually the family member petitioning for their admission). Id. § 1182(a)(4)(C) (rendering a family-sponsored noncitizen “in- admissible under this paragraph” unless the sponsor executes an “affidavit of support described in [8 U.S.C. § 1183a] with respect to such alien”).8 The affidavit provision had been in- serted into the INA weeks earlier by the Welfare Reform Act. See Welfare Reform Act § 423. In addition to making the affi- davit of support mandatory under the public charge provi- sion, IIRIRA significantly expanded 8 U.S.C. § 1183a by spelling out what the affidavit of support requires. The affidavit provision is meant to establish that the appli- cant “is not excludable as a public charge.” 8 U.S.C. § 1183a(a)(1). To that end, it empowers the federal govern- ment, as well as state and local governments, to demand re- imbursement from the sponsor for any means-tested public benefit received by the sponsored noncitizen. 9 Id. § 1183a(b)(1)(A). A “means-tested public benefit” is one 8 IIRIRA originally provided that a family-based applicant was “ex- cludable” without the affidavit. IIRIRA § 531(a). A subsequent amend- ment to the INA changed the terminology from “excludable” to “inadmis- sible.” 9 It also requires the sponsor “to maintain the sponsored alien at an annual income that is not less than 125 percent of the Federal poverty line.” 8 U.S.C. § 1183a(a)(1)(A). No 19-3169 61 available to those whose income falls below a certain level. The provision explicitly excludes certain benefits, regardless of whether they are means tested, from the sponsor’s reim- bursement obligation; by implication, receipt of every other means-tested benefit is included. See id. § 1183a note. 10 If the sponsor doesn’t pay upon request, the government can sue the sponsor. Id. § 1183a(b)(2). If the sponsor doesn’t keep “the Attorney General and the State in which the sponsored alien is currently a resident” apprised of any change in the spon- sor’s address, she is subject to a civil penalty—and that pen- alty is higher if she fails to update her address “with knowledge that the sponsored alien has received any means- tested public benefits” other than those described in three cross-referenced provisions of the Welfare Reform Act. Id. § 1183a(d). 11 The affidavit is generally enforceable for ten years or until the sponsored noncitizen is naturalized. Id. § 1183a(a)(2).12 10 I discuss these exemptions, which are narrow, in my analysis at Chevron step two. 11 This list of exempted benefits in the change-of-address penalty sec- tion largely track those in the “benefits subject to reimbursement” section. 12 IIRIRA contained another provision relevant to the “public charge” ground of inadmissibility: section 564 of the Act directed the Attorney General to establish a pilot program “to require aliens to post a bond in addition to the affidavit requirements under [8 U.S.C. § 1183a].” IIRIRA § 564(a)(1). The bond covered the cost of benefits described in the affidavit provision—that is, any means-tested benefit other than those described in three cross-referenced provisions of the Welfare Reform Act. Id. Congress instructed the Attorney General to set the bond at “an amount that is not less than the cost of providing [the relevant benefits] for the alien and the alien’s dependents for 6 months.” Id. § 564(b)(2). If an admitted noncitizen used a covered benefit, the government could bring suit either on the bond 62 No. 19-3169 Notwithstanding IIRIRA’s obvious—and obviously sig- nificant—amendments to the public charge provision, the plaintiffs insist, and the majority agrees, that its amendments reveal nothing about the scope of the term “public charge.” Yet as I will explain below, the 1996 amendments were not only material, but they also increased the bite of the public charge exclusion. The plaintiffs characterize the affidavit provision as hav- ing nothing to do with admissibility; as they see it, the provi- sion merely reinforces restrictions on government benefits for lawful permanent residents. They offer two basic arguments in support of that position: first, that the supporting-affidavit requirement appears in a different provision than does the public charge exclusion (8 U.S.C. § 1183a, as opposed to § 1182(a)(4)), and second, that the supporting-affidavit re- quirement doesn’t apply to everyone who is subject to the public charge exclusion. The first argument is totally unpersuasive. The public charge provision explicitly cross-references the affidavit pro- vision, thereby tying the two together, and it makes obtaining an affidavit of support a condition of admissibility. Id. § 1182(a)(4)(C)(ii). What’s more, the affidavit provision ex- pressly states that the point of the affidavit is “to establish that an alien is not excludable as a public charge under section 1182(a)(4).” Id. § 1183a(a)(1). Because a family-sponsored ap- plicant is inadmissible as a public charge without the affida- vit, the coverage of the affidavit is very strong evidence of the or against the sponsor pursuant to 8 U.S.C. § 1183a. IIRIRA § 564(a)(2). Congress allowed this pilot program to sunset after three years. Id. § 564(e). No 19-3169 63 nature of the burden with which the public charge exclusion is concerned. 13 The plaintiffs’ second argument fails too. As an initial mat- ter, the affidavit provision—which, it bears repeating, is tied by cross-reference to the public charge exclusion—uses the term “public charge,” and we “do[] not lightly assume that Congress silently attaches different meanings to the same term in the same or related statutes.” Azar v. Allina Health Servs., 139 S. Ct. 1804, 1812 (2019); see also Desert Palace, Inc. v. Costa, 539 U.S. 90, 101 (2003) (explaining that as a general rule, “identical words used in different parts of the same act are intended to have the same meaning” (citation omitted)). The plaintiffs don’t specify what different meaning the term “pub- lic charge” might have in the affidavit provision; they just vaguely assert that the provision is getting at something else. They presumably don’t want to embrace the logical implica- tion of their position: that the term “public charge” means something more stringent for family-based immigrants, who need to produce an affidavit, than it does for the others, who don’t. In any event, this argument assumes that if the affidavit were tied to the standard of admissibility, Congress would have required one from everyone subject to the exclusion. Its choice to require an affidavit only from family-based immi- grants, the logic goes, means that the affidavit provision can’t 13 The same is true of IIRIRA’s pilot bond program. The required bond protected the government against the risk that the noncitizen would be- come a public charge, so the scope of its coverage is a window into the meaning of the term at the time of the 1996 amendments. 64 No. 19-3169 shed any light on the admissibility provision, which is more generally applicable. This argument is misguided. There is an obvious explana- tion for why Congress required supporting affidavits from family-based immigrants and not from employment-based immigrants or green card lottery winners: that is the only con- text in which it makes sense to demand this assurance. A con- nection to a citizen or lawful permanent resident is the basis for a family-based green card. 8 U.S.C. §§ 1151(b)(2), 1153(a). The same is not true for immigrants who obtain diversity or employment-based green cards, neither of which is based on a personal relationship—much less a relationship close enough that someone would be willing to take on ten years’ worth of potentially significant liability. Moreover, in the con- text of an employment-based green card, a supporting affida- vit would add little. The affidavit is a means of providing the Executive with assurance that the green card applicant will not become a public charge if admitted. The stringent criteria for an employment-based green card provide similar assur- ance. Employment-based green cards are reserved largely for those with “extraordinary ability in the sciences, arts, educa- tion, business, or athletics which has been demonstrated by sustained national or international acclaim”; “outstanding professors and researchers” who are “recognized internation- ally”; “multinational executives and managers”; those who hold advanced degrees and have job offers; and entrepre- neurs prepared to invest a minimum of $1,000,000 in a ven- ture that will benefit the United States economy and employ “not fewer than 10 United States citizens or [lawful perma- nent residents].” Id. § 1153(b)(1)–(5). Someone who meets these criteria is unlikely to have trouble supporting herself in the future. That said, if an employment-based applicant will No 19-3169 65 be working for a relative, and therefore has a family connec- tion, the statute still requires her to obtain a supporting affi- davit—demonstrating that the affidavit is not uniquely appli- cable to those applying for family-based green cards. See id. § 1182(a)(4)(D). Despite the plaintiffs’ effort to show otherwise, it doesn’t make sense to treat the affidavit provision as an anomalous carve-out rather than compelling evidence of the scope of the public charge inquiry. In fact, trying to categorize the sup- porting affidavit as limited by virtue of its application to fam- ily-based immigrants is a sleight of hand, because, as the plaintiffs surely know, the family-based category is not simply one among several to which the public charge exclu- sion applies. As a practical matter, it is the category for which the exclusion matters most. The number of lottery winners is considerably smaller than the number of family-based immi- grants, and employment-based immigrants—also a smaller category than the family based—have other means of demon- strating self-sufficiency. In short, the 1996 amendments to the public charge provi- sion—most notably, the addition of factors to guide the public charge determination and the insertion of the affidavit re- quirement—were material. What’s more, the affidavit provi- sion reflects Congress’s view that the term “public charge” encompasses supplemental as well as primary dependence on public assistance. To establish that a family-based applicant is not excludable as a public charge, a sponsor must promise to pay for the noncitizen’s use of any means-tested benefit out- side the itemized exclusions. Without such an affidavit, the noncitizen is inadmissible. Congress’s attempt to aggressively 66 No. 19-3169 protect the public fisc through the supporting-affidavit re- quirement is at odds with the view that it used the term “pub- lic charge” to refer exclusively to primary and permanent de- pendence. C. Switching gears, the plaintiffs—with the support of the House of Representatives, appearing as amicus curiae—ad- vance a creative structural argument for why the term “public charge” must be interpreted narrowly: they say that interpret- ing the term to include the receipt of supplemental benefits is inconsistent with Congress’s choice in the Welfare Reform Act to make such benefits available to lawful permanent resi- dents. According to the plaintiffs, Congress would not have authorized lawful permanent residents to receive supple- mental benefits if it did not expect them to use those benefits. And it is inconsistent with Congress’s generosity to deny someone a green card because she is likely to take advantage of benefits for which Congress has made her eligible. The stat- utory scheme therefore forecloses the possibility of interpret- ing “public charge” to mean anything other than primary and permanent dependence. There are several problems with this argument. To begin with, its logic would read the public charge provision out of the statute. The premise of the public charge inquiry has al- ways been that immigrants in need of assistance would have access to it after their arrival—initially through state poor laws and later through modern state and federal welfare sys- tems. Indeed, it is difficult to imagine how someone could be- come a public charge under any conception of the term if it were impossible to receive public aid. For example, on the plaintiffs’ logic, DHS could not exclude an applicant even if it No 19-3169 67 predicted that the applicant would eventually become perma- nently reliant on government benefits, because the future use of those benefits would, after all, be authorized. Barring the Executive from considering a green card applicant’s potential use of authorized benefits would render the statutory public charge exclusion a dead letter. Moreover, the plaintiffs’ position assumes that tension ex- ists between the public charge exclusion and the availability of benefits to lawful permanent residents—and that this ten- sion can be resolved only by limiting the scope of the exclu- sion. In fact, the public charge exclusion and the availability of benefits are easily reconcilable. Immigration law has long distinguished between one who becomes a public charge be- cause of a condition preexisting her arrival and one who be- comes a public charge because of something that has hap- pened since. See, e.g., id. § 1227(a)(5) (“Any alien who, within five years after the date of entry, has become a public charge from causes not affirmatively shown to have arisen since en- try is deportable.”); Act of Mar. 3, 1891, ch. 551, § 11, 26 Stat. 1084, 1086 (”[A]ny alien who becomes a public charge within one year after his arrival in the United States from causes ex- isting prior to his landing therein shall be deemed to have come in violation of law and shall be returned as aforesaid.”). Providing benefits to immigrants who have been here for a designated period of time—generally five years under current law—takes care of immigrants in the latter situation. Life con- tains the unexpected: for instance, a pandemic may strike, leaving illness, death, and job loss in its wake. A lawful per- manent resident who falls on hard times can rely on public assistance to get back on her feet. Congress’s willingness to authorize funds to help immigrants who encounter unex- 68 No. 19-3169 pected trouble is perfectly consistent with its reluctance to ad- mit immigrants whose need for help is predictable upon arri- val. In any event, the plaintiffs’ argument is inconsistent not only with the statutory exclusion, but also with the Welfare Reform Act. As the plaintiffs tell it, Congress has generously supported noncitizens, thereby implicitly instructing the Ex- ecutive to ignore a green card applicant’s potential usage of supplemental benefits in the admissibility determination. But that is a totally implausible description of the Welfare Reform Act. The stated purpose of the Act is to ensure that noncitizens “within the Nation’s borders not depend on public resources to meet their needs, but rather rely on their own capabilities and the resources of their families, their sponsors, and private organizations,” and that “the availability of public benefits not constitute an incentive for immigration to the United States.” 8 U.S.C. § 1601(2). To this end, the Act renders lawful permanent residents ineligible for most benefits until they have lived in the United States for at least five years. Id. § 1613(a). The Act’s dramatic rollback of benefits for nonciti- zens sparked vociferous criticism. See Isabel Sawhill et al., Problems and Issues for Reauthorization, in WELFARE REFORM AND BEYOND: THE FUTURE OF THE SAFETY NET 20, 27 (Isabel Sawhill et al. eds., 2002) (referring to the five-year aid eligibil- ity restriction as one of the Act’s “most contentious features”). It blinks reality to describe the Welfare Reform Act as a “grant” of benefits, as the plaintiffs do, or to say that the Act No 19-3169 69 took an immigrant’s potential use of supplemental benefits off the table for purposes of the admissibility determination. 14 *** Given the length and complexity of my analysis of the plaintiffs’ arguments at Chevron step one, a summary may be helpful. In my view, the plaintiffs can’t show that the term “public charge” refers narrowly to someone who is primarily and permanently dependent on government assistance. The term “public charge” was broad when it entered federal im- migration law in 1882, and it has not been pinned down since. IIRIRA, Congress’s latest word on the public charge provi- sion, cuts in the opposite direction of the plaintiffs’ argument, as does the Welfare Reform Act, which, contrary to the plain- tiffs’ argument, hardly reflects a congressional desire that im- migrants take advantage of available public assistance. In fact, the amendments that IIRIRA and the Welfare Reform Act to- gether made to the INA reflect more than Congress’s view that the term “public charge” is capacious enough to include supplemental dependence on public assistance. They reflect 14 As the plaintiffs point out, Congress softened some of these re- strictions in subsequent legislation. Perhaps most notably, in 2002 Con- gress passed the Farm Security and Rural Investment Act, which made adults eligible for SNAP after 5 years of residency (it had previously been 10) and children eligible for SNAP immediately after becoming lawful permanent residents. Pub. L. No. 107-171, § 4401, 116 Stat. 134, 333 (2002) (codified as amended at 8 U.S.C. § 1612(a)(2)). Yet these minor adjust- ments, even if slightly more generous than the original restrictions, did not overhaul immigration policy—nor, as I have already explained, is it unreasonable in any event for the Executive to consider whether a green card applicant is likely to use benefits if she is permitted to stay. That’s the point of the public charge determination. 70 No. 19-3169 its preference that the Executive consider even supplemental dependence in enforcing the public charge exclusion. III. While the term “public charge” is indeterminate enough to leave room for interpretation, DHS can prevail only if its definition is reasonable. The majority holds that DHS is likely to lose on the merits of that argument; I disagree. My dissent from the majority on this score is inevitable, given how differ- ently we analyze the statute at Chevron step one. The majority seems to understand “public charge” to mean something only slightly broader than “primarily and permanently depend- ent,” but I understand it to be a much more capacious term— not only as a matter of history, but also by virtue of the 1996 amendments to the public charge provision. On my reading, in contrast to the majority’s, the statute gives DHS relatively wide discretion to specify the degree of benefit usage that ren- ders someone a “public charge.” Thus, the majority and I ap- proach Chevron step two from different starting points. The plaintiffs challenge the reasonableness of the rule’s definition in two respects. First, they object to the particular benefits that DHS has chosen to designate in its definition of “public charge.” According to the plaintiffs, DHS has unrea- sonably interpreted the statute insofar as the rule counts in- kind aid. Second, they argue that DHS has set the relevant benefit usage so low that the definition captures people who cannot reasonably be characterized as “public charges.” I will address these arguments in turn. A. The plaintiffs don’t contest DHS’s authority to account for the receipt of state and federal cash assistance (like SSI and No 19-3169 71 TANF) in the definition of “public charge.” But they insist that in-kind benefits (like SNAP, public housing, and Medi- caid) are off-limits. Their argument in support of that position is difficult to grasp. In their brief, the plaintiffs vaguely assert that in-kind benefits shouldn’t be counted because they are categorically different from cash payments; they imply that the term “public charge” does not encompass someone who relies on in-kind public assistance. At oral argument, the plaintiffs wisely abandoned that position. For one thing, they could not articulate why it mattered whether the government chose to give someone $500 for groceries or $500 worth of food. For another, that argument is inconsistent with history: everyone agrees that someone living permanently in a late nineteenth-century poorhouse qualified as a public charge, and shelter in a poorhouse is in-kind relief. At least rhetorically, a great deal of the plaintiffs’ argu- ment involves their repeated emphasis on the fact that the 1999 Guidance directed officers “not [to] place any weight on the receipt of non-cash public benefits (other than institution- alization) or the receipt of cash benefits for purposes other than for income maintenance.” 1999 Guidance, 64 Fed. Reg. at 28,689. The implication is that the 1999 Guidance reflects the only reasonable interpretation of the statute. Of course, the fact that a prior administration interpreted a statute differently does not establish that the new interpre- tation is unreasonable—the premise of Chevron step two is that more than one reasonable interpretation of the statute ex- ists. See Chevron U.S.A. Inc. v. Nat. Res. Def. Council, Inc., 467 U.S. 837, 863–64 (1984) (“An initial agency interpretation is not instantly carved in stone. On the contrary, the agency, to 72 No. 19-3169 engage in informed rulemaking, must consider varying inter- pretations and the wisdom of its policy on a continuing ba- sis.”). Moreover, the focus on cash benefits in the 1999 Guid- ance flowed from the Immigration and Naturalization Ser- vice’s decision to interpret “public charge” to mean “primar- ily dependent on the government for subsistence.” 1999 Guid- ance, 64 Fed. Reg. at 28,692. As the Guidance explained, INS had decided “that the best evidence of whether an alien is pri- marily dependent on the government for subsistence is either (i) the receipt of public cash assistance for income mainte- nance, or (ii) institutionalization for long-term care at govern- ment expense.” Id. DHS has now taken a different approach— it has decided that projected reliance on government benefits need not be primary to trigger the public charge exclusion. And once DHS made that baseline choice, a broader range of benefits became relevant. Thus, the plaintiffs’ fundamental objection to the counting of benefits like Medicaid, housing, and SNAP—that they are supplemental—is really just a re- packaging of their argument under Chevron step one. The plaintiffs also advance a legislative-inaction argu- ment: in 2013—twenty years after Congress enacted IIRIRA— the Senate Judiciary Committee, while debating the Border Security, Economic Opportunity, and Immigration Moderni- zation Act, voted down a proposal to require applicants for lawful permanent resident status “to show they were not likely to qualify even for non-cash employment supports such as Medicaid, the SNAP program, or the Children’s Health In- surance Program (CHIP).” S. REP. NO. 113-40, at 42 (2013). But the failure of this proposal is neither here nor there. As the Supreme Court has cautioned, “Congressional inaction lacks ‘persuasive significance’ because ‘several equally tenable in- No 19-3169 73 ferences’ may be drawn from such inaction, ‘including the in- ference that the existing legislation already incorporated the offered change.’” Pension Benefit Guar. Corp. v. LTV Corp., 496 U.S. 633, 650 (1990) (citation omitted). This rejected pro- posal—which would have overridden the 1999 Guidance—is a case in point: the rejection is as consistent with the choice to leave the matter within the Executive’s discretion as it is with the choice to force the Executive’s hand. The plaintiffs’ argu- ment has other problems too. Why should the views of the 2013 Senate Judiciary Committee be attributed to Congress as a whole? See Thompson v. Thompson, 484 U.S. 174, 191–92 (1988) (Scalia, J., concurring in the judgment) (“Committee re- ports, floor speeches, and even colloquies between Congress- men, are frail substitutes for bicameral vote upon the text of a law and its presentment to the President.” (citation omitted)). And how could the unenacted views of the 2013 Congress set- tle the meaning of language chosen by a different Congress at a different time? See United States v. Price, 361 U.S. 304, 313 (1960) (“[T]he views of a subsequent Congress form a hazard- ous basis for inferring the intent of an earlier one.”). Thus, the plaintiffs are wrong to insist that DHS is barred from considering the receipt of a particular benefit simply be- cause the benefit is in-kind rather than cash. There is no such bar. Rather, the list of designated benefits is reasonable if re- ceiving them is consistent with the lack of self-sufficiency con- veyed by the term “public charge.” Answering this question requires fleshing out what it means to lack self-sufficiency for purposes of the public charge exclusion. As the majority observes, no one is self-suf- ficient in an “absolutist” sense because everyone relies on 74 No. 19-3169 some nonmonetary government services—for example, pub- lic snow removal and emergency services. Maj. Op. at 13, 37. Importantly, the term “public charge” does not implicate self- sufficiency in this absolutist sense. Throughout its centuries- long history, “public charge” has always been associated with dependence on a particular category of government pro- grams: those available based on financial need. In the nine- teenth and early twentieth centuries, these were “poor relief” programs; now, they are the need-based programs of the modern welfare system. And what has always been implicit in the term “public charge” was made explicit by the 1996 amendments. The statutory exclusion requires the Executive to consider the noncitizen’s age; health; family status; assets, resources, and financial status; and education and skills—fac- tors plainly designed to determine whether a noncitizen will be able to support herself, not whether she will use generally available services like snow removal. In the same vein, the sponsor’s reimbursement obligation covers only those bene- fits that are “means tested”—that is, available to those whose income falls below a certain threshold. As a matter of both history and text, a “public charge” lacks self-sufficiency in the sense that she lacks the financial resources to provide for her- self. The benefits designated in DHS’s definition are all con- sistent with this concept of self-sufficiency. Recall that DHS has designated the following benefits: cash assistance for in- come maintenance (including SSI, TANF, and state cash assis- tance), SNAP, the Section 8 Housing Choice Voucher Pro- gram, Section 8 project-based rental assistance, housing ben- efits under Section 9, and Medicaid (with some explicit excep- tions). 8 C.F.R. § 212.21. These benefits are all means tested; No 19-3169 75 they are also squarely within the Welfare Reform Act’s defi- nition of “public benefit.” 8 U.S.C. §§ 1611(c), 1621(c) (defin- ing “public benefit” to include welfare, food, health, and pub- lic-housing benefits funded by the federal, state, or local gov- ernments). It is consistent with the term “public charge” to consider the potential receipt of cash, food, housing, and healthcare benefits—all of which fulfill fundamental needs— in evaluating whether someone is likely to depend on public assistance to get by. It is also worth noting some of the benefits that the rule does not include: significantly, the rule’s definition accommo- dates the reimbursement limitations in the affidavit provi- sion. Under the affidavit provision, the following benefits, even if means tested, are not subject to reimbursement: certain forms of emergency medical assistance; short-term, in-kind emergency disaster relief; school-lunch benefits; benefits un- der the Child Nutrition Act of 1966; public-health assistance for immunization, as well as treatment for the symptoms of communicable disease; certain foster-care and adoption pay- ments; certain in-kind services such as soup kitchens and cri- sis counseling; student assistance for higher education; bene- fits under the Head Start Act; means-tested programs under the Elementary and Secondary Education Act of 1965; and certain job-training benefits. Id. § 1183a note.15 These exemptions under the affidavit provision are ex- cluded from the rule too. The rule’s definition provides “an 15 By virtue of a notice issued by the Department of Housing and Ur- ban Development, housing benefits are excluded from the reimbursement obligation. See 8 C.F.R. § 213a.1; Eligibility Restrictions on Noncitizens, 65 Fed. Reg. 49,994 (Aug. 16, 2000). But that exemption is not statutory, and here, I’m concerned only with DHS’s interpretation of the statute. 76 No. 19-3169 exhaustive list of public benefits,” Inadmissibility on Public Charge Grounds, 84 Fed. Reg. at 41,296, so any benefit not mentioned in the list is by implication excluded from the def- inition. And the list does not mention any of the benefits ex- empted in the affidavit provision of the statute. 8 C.F.R. § 212.21; see also Inadmissibility on Public Charge Grounds, 84 Fed. Reg. at 41,312 (noting that the rule’s “definition does not include benefits related exclusively to emergency response, immunization, education, or social services”); id. at 41,482 (ex- plaining that the rule’s definition “does not include emer- gency aid, emergency medical assistance, or disaster relief”); id. at 41,389 (excluding benefits under the National School Lunch Act, the Child Nutrition Act, and the Head Start Act). Indeed, to highlight just how carefully the rule tracks the stat- utory exemptions to the affidavit of support, consider the rule’s exclusion of Medicaid for those under the age of 21 and pregnant women. Id. at 41,367. These benefits do not appear in the list of exemptions to the affidavit of support, but they are exempted from the sponsor’s reimbursement obligations under a different statutory provision. 42 U.S.C. § 1396b(v)(4)(B). The rule captures that exclusion even though it appears elsewhere; in other words, DHS did not simply copy and paste the statutory note. In sum, the designated benefits are not only consistent with the term “public charge,” but they also fit neatly within the statutory structure. Considering the potential receipt of these benefits to gauge the likelihood that a noncitizen will become a public charge is therefore not an unreasonable in- terpretation of the statute. No 19-3169 77 B. The closer question is whether DHS’s benefit-usage threshold stretches the meaning of “public charge” beyond the breaking point. The rule defines “public charge” to mean a noncitizen who receives one or more of the designated ben- efits “for more than twelve months in the aggregate within any 36-month period.” One month of one benefit counts to- ward the twelve. As a result, an applicant expected to live in Section 8 housing for a year would be denied admission as someone who is likely to become a public charge, as would an applicant who is expected to receive three months’ worth of housing, TANF, Medicaid, and SNAP. The plaintiffs have a legislative-inaction argument for this feature of the rule too. They point out that during the enact- ment of IIRIRA, the Senate Judiciary Committee, while nego- tiating the House-passed version of the bill, dropped lan- guage that “would have clarified the definition of ‘public charge’” in the deportation provision to provide for deporta- tion if a noncitizen “received Federal public benefits for an aggregate of 12 months over a period of 7 years.” 142 Cong. Rec. S11,872, S11,882 (daily ed. Sept. 30, 1996) (statement of Sen. Kyl). Thus, they say, Congress has foreclosed the possi- bility that 12 months’ worth of benefit usage renders someone a public charge. Whatever the statutory floor is, it must be higher than that. I’ve already identified some of the problems with legisla- tive-inaction arguments, so I won’t belabor them here. It’s worth noting, though, that this legislative-inaction argument is even worse than the plaintiffs’ other. So far as the plaintiffs’ citation reveals, the proposal dropped out of the statute in the course of committee negotiations, not by a vote, and there is 78 No. 19-3169 no explanation for why it did. See Thompson, 484 U.S. at 191 (Scalia, J., concurring in the judgment) (“An enactment by im- plication cannot realistically be regarded as the product of the difficult lawmaking process our Constitution has pre- scribed.”). Moreover, the dropped proposal involved the pub- lic charge deportation provision, not the public charge admissi- bility provision. See 8 U.S.C. § 1227(a)(5). Drawing general conclusions from a committee’s decision to drop this lan- guage in a context with much higher stakes is a particularly dubious proposition. Despite the plaintiffs’ effort to demon- strate otherwise, the statute doesn’t draw a bright line requir- ing something more than 12 months of benefit usage to meet the definition of “public charge.” At oral argument, DHS declined to identify any limit to its discretion, implying that it could define public charge to in- clude someone who took any amount of benefits, no matter how small. It may have been grounding its theory in the affi- davit provision, which triggers the sponsor’s liability once the noncitizen receives “any means-tested public benefit” that falls within the sponsor’s reimbursement obligation. Id. § 1183a(b)(1)(A) (emphasis added). That may well overread the affidavit provision, which does not purport to define “public charge.” Enforcement of the public charge exclusion has waxed and waned over time in response to economic conditions, immigration policy, and changes in the programs available to support the poor. The amendments made by IIRIRA and the Welfare Reform Act, including the affidavit provision, reflect Congress’s interest in vigorous enforcement. Yet Congress left the centuries-old term in the statute, and that term has always been associated No 19-3169 79 with a lack of self-sufficiency. So that’s the principle that gov- erns here: if it’s reasonable to describe someone who takes one or more of the designated benefits “for more than twelve months in the aggregate within any 36-month period” as lack- ing in self-sufficiency, then DHS’s definition falls within the permissible range. In deciding this question, it is wrong to focus exclusively on the durational requirement—duration must be viewed in the context of the benefits measured. Three features are par- ticularly important in this regard: the designated benefits are means tested, satisfy basic necessities, and are major welfare grants. To see the importance of these features, consider how different the durational threshold would look without them—for example, if the rule measured the usage of benefits that are not means tested (e.g., public education), that are means tested but don’t satisfy a basic necessity (e.g., Pell grants), or that satisfy a basic necessity but are not major wel- fare grants (e.g., need-based emergency food assistance). Re- lying on the government to provide a year’s worth of a basic necessity (food, shelter, medicine, or cash assistance for in- come maintenance) implicates self-sufficiency in a way that funding a year of college with the help of a Pell grant does not. The plaintiffs particularly object to the rule’s stacking mechanism, which can reduce the durational requirement from 12 months to as little as 3 months. But here, too, the con- text matters: all of the designated benefits supply basic neces- sities, and the reduction is triggered in proportion to the de- gree of reliance on the government. The more supplemental the reliance, the longer it can go on before crossing the “public charge” threshold. The briefest durational threshold—three 80 No. 19-3169 months of benefit usage—meets the definition only when the recipient relies on the government for all basic necessities (food, shelter, medicine, and cash assistance for income maintenance). In other words, such short-term reliance only counts if it’s virtually total. The rule measures self-sufficiency along a sliding scale rather than by time alone. It is not unreasonable to describe someone who relies on the government to satisfy a basic necessity for a year, or mul- tiple basic necessities for a period of months, as falling within the definition of a term that denotes a lack of self-sufficiency. To be sure, the rule reaches dependence that is supplemental and temporary rather than primary and permanent. But the definition of “public charge” is elastic enough to permit that. The rule’s definition is exacting, and DHS could have exer- cised its discretion differently. The line that DHS chose to draw, however, does not exceed what the statutory term will bear. IV. This case involves more than the definition of “public charge.” The plaintiffs raised a host of objections to the rule in their complaint, and the majority addresses some of them. It concludes that the plaintiffs are likely to succeed in their chal- lenge to the factors that DHS uses to implement its definition (the list of factors includes health, family size, and English proficiency), as well as in their argument that the rule is arbi- trary and capricious. See 5 U.S.C. § 706(2)(A); Motor Vehicle Mfrs. Ass'n of the U.S., Inc. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983) (explaining that the agency must “articulate a satisfactory explanation for its action including a ‘rational connection between the facts found and the choice made’” (ci- tation omitted)). No 19-3169 81 I wouldn’t reach these issues. The district court didn’t ad- dress them, and on appeal, the parties devoted their briefs al- most entirely to the definition of “public charge.” Singleton v. Wulff, 428 U.S. 106, 120 (1976) (“It is the general rule, of course, that a federal appellate court does not consider an issue not passed upon below.”); see also Ctr. for Individual Freedom v. Van Hollen, 694 F.3d 108, 111 (D.C. Cir. 2012) (remanding to the district court for arbitrary-and-capricious review when the district court resolved a case at Chevron step one without reaching the issue and when the agency’s position was not well developed). And while it’s generally prudent to refrain from deciding difficult issues without the benefit of argu- ments from the parties, the procedural posture of this case of- fers a particularly good reason to stop where the parties did. We are reviewing the issuance of the “extraordinary remedy” of a preliminary injunction. Whitaker ex rel. Whitaker v. Kenosha Unified Sch. Dist. No. 1 Bd. of Educ., 858 F.3d 1034, 1044 (7th Cir. 2017). Based on the record developed thus far, the plain- tiffs have not shown that they are entitled to this extraordi- nary remedy. I would remand so that the district court can assess whether the plaintiffs’ remaining challenges to the rule are likely to succeed. *** The many critics of the “public charge” definition charac- terize it as too harsh. But the same can be said—and has been said—of IIRIRA and the Welfare Reform Act. The latter dra- matically rolled back the availability of aid to noncitizens, and both statutes linked those cuts to the public charge provision by making the affidavit of support a condition of admissibil- ity. The definition in the 1999 Guidance tried to blunt the force of these changes; now, DHS has chosen to exercise the leeway 82 No. 19-3169 that Congress gave it. At bottom, the plaintiffs’ objections re- flect disagreement with this policy choice and even the statu- tory exclusion itself. Litigation is not the vehicle for resolving policy disputes. Because I think that DHS’s definition is a rea- sonable interpretation of the statutory term “public charge,” I respectfully dissent.