Karen VAUGHN, et al., Plaintiffs-Appellants, v. Cheryl SULLIVAN, et al., Defendants-Appellees.
Means-tested public assistance programs place a tax on earnings. Not a direct tax, after the fashion of the Internal Revenue Code, but an indirect one. Greater earnings yield less assistance. This is what it means to say that a program is means-tested, with benefits concentrated on persons with lower incomes or wealth. At what rate should extra earnings reduce the levels of assistance? If every dollar of earnings reduces public assistance by a dollar, then beneficiaries have little or no incentive to work, may even have a strong incentive not to work (for work not only requires effort but also exposes the worker to income tax; the effective tax rate can exceed 100 percent when explicit and implicit taxes are counted). Recipients who lack an incentive to work fall behind in the marketplace; skills do not keep up; and then even if the program's tax rate changes, to encourage employment, it may be too late. But if every dollar of earnings reduces benefits only by a little (say 10 percent), in order to reward employment and encourage the transition from public assistance to self-sufficiency, the cost of the program balloons. Suppose a program guarantees everyone a minimum of $5,000 per year in cash or benefits and imposes a tax rate of 10 percent. Then even someone who earns $40,000 per year will still receive $1,000 under the program. Public assistance programs would collapse if everyone who earned less than $50,000 a year were a recipient of net transfers. Those who earn more than $50,000 would pay a staggering rate of tax to finance the program-if it could be financed at all-and the disincentive to work would affect society's most productive members.
One way to reduce a high break-even point between net transfers and net taxes ($50,000 in this example) is to reduce the guaranteed minimum, but a low minimum may defeat the purpose of the program (to provide decent housing, basic nutrition, or essential medical care). Another is to increase the implicit tax on benefits, which reinforces the cycle of dependence. Still another approach is to vary the tax rate-to have a low rate for a time (to encourage work) followed by a high rate later. This produces a notch in the relation between income and benefits. After earnings exceed some threshold, for some period of time, public benefits vanish (or are greatly diminished). The effective tax rate at the notch is huge (1 cent of additional earnings can cause the loss of a $5,000 package of benefits), and the existence of this precipice can cause people to keep their earnings under the threshold, but some participants may be able to land jobs so remunerative that they willingly forego all benefits and make the transition to self-sufficiency.
This is the premise of the Plan for Achieving Self-Support (PASS) program, see 42 U.S.C. §§ 1382a(b)(4)(B)(iv), 1382b(a)(4), under which earnings from new employment may be disregarded for as long as four years in computing income for purposes of the Supplemental Security Income (SSI) program, a state-federal cooperative means-tested program. Congress and the states that have elected to participate in the PASS program hope that people achieve levels of earnings that will induce them to surrender their SSI benefits when the period of exclusion ends. When this happens, the short-run cost to the SSI program of maintaining full benefits for people with increased earnings is swamped by the long-run savings from terminating SSI payments for future years (and by the taxes collected on the former recipients' earnings). There is, of course, another possibility: when the exclusion period ends, recipients will quit their jobs (or scale back their work) to protect their benefits. When that occurs, the PASS program not only fails to achieve its objective but also increases the cost of the SSI program (by foregoing the implicit tax on the earnings in the interim). Which effect predominates is an interesting empirical question, which may differ across classes of recipients. For example, the chronically ill may have a lower success rate than those who filed claims to SSI benefits because of a decline in local employment opportunities.
Eligibility for SSI benefits potentially affects other benefits too. As a rule, for example, anyone who is eligible for SSI benefits also is eligible for Medicaid benefits-and, for many recipients, the cost to the state of medical care delivered under Medicaid substantially exceeds the cash transfers the state provides under the SSI program. States accordingly are tempted to break the link between SSI and Medicaid, providing the cheaper SSI payments while withholding or limiting Medicaid benefits. Federal law prevents states from yielding to this temptation-unless the state in question is a “sec. 209(b) state.” When Congress linked the SSI and Medicaid programs in 1971, it feared that states would find Medicaid costs prohibitive and withdraw from the SSI program in order to limit their outlays. To avoid this, Congress gave states the option to treat eligibility for SSI and Medicaid programs separately, provided the state's benefits were no less generous than those the state provided on January 1, 1972. This option, extended by § 209(b) of a statute whose name no longer matters, is codified in 42 U.S.C. § 1396a(f). See generally Schweiker v. Hogan, 457 U.S. 569, 102 S.Ct. 2597, 73 L.Ed.2d 227 (1982); Herweg v. Ray, 455 U.S. 265, 102 S.Ct. 1059, 71 L.Ed.2d 137 (1982); Schweiker v. Gray Panthers, 453 U.S. 34, 101 S.Ct. 2633, 69 L.Ed.2d 460 (1981); Mattingly v. Heckler, 784 F.2d 258 (7th Cir.1986). Indiana took advantage of the § 209(b) option and uses different formulas to determine eligibility for, and the amount of, SSI and Medicaid benefits. One difference is under attack in this litigation: Indiana disregards earnings under a PASS program when determining the Medicaid benefits of blind persons, but it counts these earnings for everyone else. 405 Ind. Admin. Code §§ 2-3-3(8), 2-3-15(c)(9).
Matthew Ravin and Karen Vaughn, the two representative plaintiffs in this class action, are severely disabled. Ravin, who has multiple sclerosis, cannot control his legs or use his hands well enough to feed himself. Vaughn is a quadriplegic, paralyzed below the neck as a result of a shooting accident. Like other members of the plaintiff class, Ravin and Vaughn incur heavy medical expenses that have exhausted their resources; they receive Medicaid benefits. Vaughn and Raven have personal PASS programs approved, for purposes of the SSI program, by the Social Security Administration. But because they are not blind, Indiana treats half of their earnings in excess of a threshold as available for the payment of medical bills; it also requires them to devote all liquid assets exceeding $1,500 to medical care before the state will cover remaining expenses. See 405 Ind. Admin. Code §§ 2-3-3 through 2-3-10. They are somewhat better off with the PASS program than without; the PASS program enhances the SSI payments they receive at a given level of income from employment, and Indiana does not require them to use current SSI income for medical care (though it did count a lump-sum payment of back SSI benefits against Raven's resource limit). But they would be better off still if employment earnings were disregarded for Medicaid purposes-as the earnings of blind persons under PASS programs are disregarded. Raven wants to save $9,000 from his PASS earnings to buy a handicapped-accessible van, but Indiana would kick him out of the Medicaid program as soon as his savings exceed $1,500. Plaintiffs contend that Indiana's more favorable treatment of blind persons violates two federal statutes and the equal protection clause of the fourteenth amendment. The district court granted Indiana's motion for summary judgment. 906 F.Supp. 466 (S.D.Ind.1995).
Like the district court, we start with the argument that Indiana has violated the Medicaid statute-in particular, 42 U.S.C. § 1396a(a)(17), which provides that every state plan must “include reasonable standards (which shall be comparable for all groups ․ ) for determining eligibility for and the extent of medical assistance under the plan”. The district court held that this requirement does not apply to Indiana, because § 209(b) negates “any other provision of this subchapter” except for a short list, on which § 1396a(a)(17) does not appear. Section 209(b) and § 1396a(a)(17) are part of the same subchapter (indeed, the same section) of the United States Code. Given the rule that unambiguous statutory text prevails over arguments based on statutory purpose and history, that ought to be the end of matters.
Plaintiffs say that it is not. They appeal to legislative purpose. Congress enacted § 209(b) to break the link between SSI and Medicaid, a link that appears in § 1396a(a)(10)(A). This means, plaintiffs insist, that the sole effect of § 209(b) is to entitle Indiana to disregard § 1396a(a)(10)(A). Section 209(b) states cannot adopt Medicaid plans that would have been unlawful in 1971, see Norman v. St. Clair, 610 F.2d 1228, 1235 (5th Cir.1980), and a version of § 1396a(a)(17) has been in the Medicaid statute since that program's creation in 1965. Plaintiffs point to several cases that have entertained arguments that one or another § 209(b) state was out of compliance with § 1396a(a)(17). E.g., Gray Panthers; Mattingly; Roloff v. Sullivan, 975 F.2d 333 (7th Cir.1992). These cases rejected the “comparability” claims on the merits, which leads plaintiffs to argue that they implicitly hold § 1396a(a)(17) applicable to § 209(b) states.
Indiana responds that “implicit” means that there was no “holding”; the opinions could bypass the effect of § 209(b) because it turned out not to matter. One may say that the purpose of § 209(b) was to cancel § 1396a(a)(10)(A), but statutes often are written more broadly than their immediate objectives, in order to avoid the risk that narrow language will undermine the objective in ways not yet contemplated. Cf. Rodriguez v. United States, 480 U.S. 522, 525-26, 107 S.Ct. 1391, 1393, 94 L.Ed.2d 533 (1987). Contentions of the kind plaintiffs present illustrate that risk. Indiana is free to use different eligibility and resource standards for the SSI and Medicaid programs; that is what § 209(b) is about; but if plaintiffs are right, the “comparability” requirement of § 1396a(a)(17) withdraws much if not all of that freedom once Indiana grants Medicaid benefits to a single person-not only because almost any variance in the standards can be said to treat “comparable” cases differently, but also because it transfers decisionmaking authority from the state to the federal judiciary. Congress required equality in deductions elsewhere in the web of social welfare programs, see 42 U.S.C. § 1309, and we take the plaintiffs' silence about that statute as a concession of its inapplicability to this dispute. Would it not be strange to introduce the same rule through the back door, when it is barred at the front by § 209(b)? The course most compatible with the Supreme Court's instructions about statutory interpretation is to take § 209(b) literally, covering the entire subchapter except for listed exclusions (of which § 1396a(a)(17) is not one).
Even if this is wrong, the second part of plaintiffs' argument does not follow. They concede that Indiana may maintain any distinction between Medicaid and SSI that was lawful in 1971. Back in 1971, Indiana practiced the same distinction: it allowed blind persons, but not other disabled persons, to disregard PASS income when calculating eligibility for Medicaid benefits. The Secretary of Health and Human Services has approved this distinction for almost three decades. (Each state needs the Secretary's approval for its program; see 42 U.S.C. § 1396a(b).) “Comparable” is a fuzzy term, as we remarked in K.R. v. Anderson Community School Corp., 81 F.3d 673 (7th Cir. 1996), when considering another program that drew distinctions on the basis of physical conditions. Agencies must have leeway when dealing with such open-ended words. “Comparable” cannot mean “identical,” for that would make § 209(b) illusory; but if states can comply with § 1396a(a)(17) without using identical standards, how great may the difference be? That's a question for the responsible administrator, in this case the Secretary. In what dimensions must standards be “comparable?” Plaintiffs answer “with respect to income and medical needs,” but again that bids fair to eviscerate § 209(b). Other dimensions may be relevant-perhaps the reason the person requires assistance (e.g., permanent disability versus transitional unemployment) and the prospect that equal treatment would produce equal results. Because the objective of a PASS program is self-sufficiency, a state might differentiate according to the likelihood of achieving that goal. A state might conclude that the blind have greater employment opportunities than do quadriplegics, coupled with lower medical costs. No one will leave the SSI and Medicaid programs unless employment income exceeds the benefits foregone, and a state may conclude that this is more likely to occur for blind persons than for other participants in the program, and therefore represents a better investment of limited funds (recall that PASS programs impose short-term costs in the form of the foregone implicit tax on earnings). Perhaps this is why the Secretary has approved Indiana's program year in and year out.
Indeed, Congress itself may have expressed a similar judgment-and in § 1396a(a)(17), no less. Here is the full text of that subsection:
[A state plan must] except as provided in subsections (l )(3), (m)(3), and (m)(4), include reasonable standards (which shall be comparable for all groups and may, in accordance with standards prescribed by the Secretary, differ with respect to income levels, but only in the case of applicants or recipients of assistance under the plan who are not receiving aid or assistance under any plan of the State approved under title I, X, XIV, or XVI, or part A of title IV, and with respect to whom supplemental security income benefits are not being paid under title XVI, based on the variations between shelter costs in urban areas and in rural areas) for determining eligibility for and the extent of medical assistance under the plan which (A) are consistent with the objectives of this title, (B) provide for taking into account only such income and resources as are, as determined in accordance with standards prescribed by the Secretary, available to the applicant or recipient and (in the case of any applicant or recipient who would, except for income and resources, be eligible for aid or assistance in the form of money payments under any plan of the State approved under title I, X, XIV, or XVI, or part A of title IV, or to have paid with respect to him supplemental security income benefits under title XVI) as would not be disregarded (or set aside for future needs) in determining his eligibility for such aid, assistance, or benefits, (C) provide for reasonable evaluation of any such income or resources, and (D) do not take into account the financial responsibility of any individual for any applicant or recipient of assistance under the plan unless such applicant or recipient is such individual's spouse or such individual's child who is under 21 or (with respect to States eligible to participate in the State program established under title XVI), is blind or permanently and totally disabled, or is blind or disabled as defined in section 1614 [42 U.S.C. § 1382c] (with respect to States which are not eligible to participate in such program); and provide for flexibility in the application of such standards with respect to income by taking into account, except to the extent prescribed by the Secretary, the costs (whether in the form of insurance premiums, payments made to the State under section 1903(f)(2)(B) [42 U.S.C. § 1396b(f)(2)(B) ], or otherwise and regardless of whether such costs are reimbursed under another public program of the State or political subdivision thereof) incurred for medical care or for any other type of remedial care recognized under State law[.]
This is hardly a straightforward requirement that eligibility standards be comparable. It is a detailed prescription for how similar, and in what respects. One distinction recognized by this language is between the blind and the disabled. Twice the subsection uses a variation on the formula “blind or disabled”, which implies that Congress sees these two conditions as different, just as Indiana does. Section 1396a(a)(17) is hardly special. The Internal Revenue Code treats blindness differently from other disabilities, and in some 30 other places the United States Code draws this distinction. The Code of Federal Regulations does so in more than 600 sections, including many concerned with Medicaid. Plaintiffs have not provided any evidence that for purposes of the PASS program “blind” and “disabled” are identical. At oral argument, plaintiffs' capable lawyer conceded that he had no idea whether, in non-sec. 209(b) states, PASS programs are as effective for members of their class as they are for the blind. There is accordingly no foundation for a conclusion that the Secretary acted arbitrarily in approving Indiana's plan.
Argument No. 2 is that Indiana's plan violates the Rehabilitation Act, 29 U.S.C. § 794, and the Americans with Disabilities Act (ada), 42 U.S.C. §§ 12101-213, because it excludes sighted disabled persons from the PASS program on account of their disability. Even if they have PASS programs approved by the federal government, they can't keep as much of their earnings as blind persons do. Like the district court, we find this a strained argument. Indiana does not exclude disabled persons from Medicaid benefits; to the contrary, disabled persons are the recipients of vast transfers in their favor, and so far as we can see the net transfers are far greater for members of the plaintiff class (who require expensive medical care and often full-time personal assistants) than for the blind (most of whom are self-sufficient). Neither the Rehabilitation Act nor the ada requires states to provide identical benefits; they retain “substantial discretion to choose the proper mix of amount, scope and duration limitations on coverage.” Alexander v. Choate, 469 U.S. 287, 303, 105 S.Ct. 712, 721, 83 L.Ed.2d 661 (1985). (Although the ada was enacted after Choate, plaintiffs acknowledge that it is not substantively different from the Rehabilitation Act with respect to questions such as this.) Plaintiffs' minimum obligation therefore is to show that the total package of benefits they receive is lower than that of some (dare we say it) comparable group. They have not tried to do so.
Even proof of such a differential would not be enough in light of Traynor v. Turnage, 485 U.S. 535, 108 S.Ct. 1372, 99 L.Ed.2d 618 (1988). Plaintiffs' understanding of the Rehabilitation Act and the ada implies that these laws require states to extend Medicaid benefits to everyone of the same real income level; any difference would discriminate impermissibly against someone with a disability (that is, with a medical condition for which Medicaid would pay). That would pretty much abolish the differences permitted by § 209(b). Traynor considered a similar argument: that in light of the Rehabilitation Act, the Veterans' Administration could not distinguish between medical benefits provided to alcoholics and those provided to persons with other disabilities. The Court replied that the Rehabilitation Act did not repeal the statute on which the VA relied, and that repeals by implication require a convincing demonstration of incompatibility. Just so here. Neither the Rehabilitation Act nor the ada repeals § 209(b) or forbids all distinctions in public welfare benefits based on physical condition. The Medicaid program is rife with these distinctions, which, the Court held in Choate, are compatible with the Rehabilitation Act. Traynor added: “There is nothing in the Rehabilitation Act that requires that any benefits extended to one category of handicapped persons also be extended to all other categories of handicapped persons.” 485 U.S. at 549, 108 S.Ct. at 1382. It follows that distinctions permitted by § 209(b) were not made unlawful by the Rehabilitation Act or the ADA.
Finally, plaintiffs turn to the equal protection clause. To provide different levels of medical benefits to people with the same real income is irrational, plaintiffs believe. Yet a legislative decision “may be based on rational speculation unsupported by evidence or empirical data.” FCC v. Beach Communications, Inc., 508 U.S. 307, 315, 113 S.Ct. 2096, 2102, 124 L.Ed.2d 211 (1993). See also, e.g., Minnesota v. Clover Leaf Creamery Co., 449 U.S. 456, 463-65, 101 S.Ct. 715, 723-24, 66 L.Ed.2d 659 (1981); Vance v. Bradley, 440 U.S. 93, 111, 99 S.Ct. 939, 949-50, 59 L.Ed.2d 171 (1979); National Paint & Coatings Ass'n v. Chicago, 45 F.3d 1124, 1127-30 (7th Cir.1995). We have discussed one potential difference between the blind and the members of the plaintiff class: a state rationally may conclude that blind participants in PASS programs are more likely to become self-sufficient. Perhaps that is wrong, though plaintiffs have not offered a scrap of data to undermine it. Wrong or not, however, such a belief could not be called irrational. States may decide to invest their funds where the returns are greatest, and they may conclude that investment in a PASS program is more likely to pay off for blind persons than for others. Cases such as Califano v. Jobst, 434 U.S. 47, 98 S.Ct. 95, 54 L.Ed.2d 228 (1977), rebuff constitutional challenges to rules in disability-benefit programs that have far weaker empirical support. In the end, the only real question is whether Indiana can persuade the Secretary to approve the way it implements the Medicaid program. She did.
EASTERBROOK, Circuit Judge.