Roundtree v. Berger , 420 F. Supp. 282 ( 1976 )

420 F. Supp. 282 (1976)

Levan ROUNDTREE and Delores Roundtree, Individually, on behalf of their minor children and all other persons similarly situated, Plaintiffs,
and
John Folsom et al.
v.
Stephen BERGER, Individually and as Acting Commissioner of the New York State Department of Social Services, et al., Defendants.

No. 75-C-1052.

United States District Court, E. D. New York.

September 17, 1976.

John C. Gray, Jr., Brooklyn, N. Y., Lloyd E. Constantine, Brooklyn, N. Y., of counsel, Brooklyn Legal Services Corp. B, for plaintiffs and plaintiff-intervenors.

Louis J. Lefkowitz, Atty. Gen., State of New York, New York City, for defendants Berger and the Dept. of Social Services; Stanley L. Kantor, Asst. Atty. Gen., of counsel.

W. Bernard Richland, Corp. Counsel, City of New York, New York City, for James Dumpson; James S. Strauss, Asst. Corp. Counsel, New York City, of counsel.

Before MESKILL, Circuit Judge, and WEINSTEIN and COSTANTINO, District Judges.

MEMORANDUM and ORDER

COSTANTINO, District Judge.

In this action, plaintiffs seek (1) a declaratory judgment voiding New York State Social Services Law § 131-i and 18 N.Y.C. R.R. § 325.19(c) as unconstitutional, and (2) an injunction restraining defendants from enforcing the provisions of these sections.

Both provisions are part of New York State's Home Relief Program. The program is completely state funded and provides cash assistance to persons who are ineligible for federally funded programs. Under the program, a standard of monthly need is determined according to the number of persons in a household. Social Services Law § 131-a(2).

If no member of the household is employed, the household receives a grant equal to the standard of need. If a member of the household is employed the household receives a cash allowance to supplement the earnings. The allowance is equal to the difference between the standard of need and the net earned income of the household after deduction of work-related expenses. These expenses include (1) all non-personal work expenses such as union dues, costs of tools, materials, uniforms and other special clothing; (2) all personal work expenses such as Federal, State and local taxes, group insurance, meals and transportation; and (3) an allowance of $20 per month as a "special work expense." However, under the provisions challenged in this lawsuit (Social Services Law § 131-i and 18 N.Y.C. R.R. § 352.19(c)) the maximum which can be deducted from gross salary for these work-related expenses is $80 per month.

The Roundtrees represent a class of people with work-related expenses of more than $80 whose net income would result in eligibility for home relief, but for the statutory relief. The Folsoms and Pelligrinos have work-related expenses of more than $80. While they are eligible for home relief aid, they receive "reduced" benefits because of the $80 limit. It is claimed that because of the $80 limit plaintiffs are denied equal protection and are deprived of their property rights without due process of law. We disagree.

Equal Protection

Although plaintiffs dispute the continuing viability of the "two-tiered" equal protection test, the Supreme Court has only recently re-affirmed its adherence to the traditional analysis. Massachusetts Board of Retirement v. Murgia, ___ U.S. ___, 96 S. Ct. 2562, 49 L. Ed. 2d 520 (1976). Since the class of those working poor who have more than $80 a month in work expenses does not constitute a suspect class, we conclude that the state classification should be examined under the rational basis standard. This standard requires that the "legislative classification must be sustained unless it is ``patently arbitrary' and bears no rational relationship to a legitimate governmental interest." Frontiero v. Richardson, 411 U.S. 677, 683, 93 S. Ct. 1764, 1768, 36 L. Ed. 2d 583 (1973). So long as the classification does bear a rational relationship to a legitimate governmental interest, the constitutionality of the statute's discrimination will be presumed "unless [it] trammels fundamental personal rights or is drawn upon inherently suspect distinctions such as race, religion or alienage." City of New Orleans v. Dukes, ___ U.S. ___, 96 S. Ct. 2513, 2516, 49 L. Ed. 2d 511 (1976).

Plaintiffs have introduced voluminous studies of related programs contending that these studies indicate that the maximum limit on work expenses set by New York State serves as a work disincentive and in fact contributes to family disintegration. These studies, however, are not relevant to our analysis. For, as plaintiffs themselves admit, "[t]he State is not bound to adopt a . . . scheme to encourage employment, it need not encourage employment at all. The State is not under an obligation to have a Home Relief program at all. . . .."^[1]

In the absence of invidious discrimination it is not the function of the judiciary to weigh or balance the side effects of legislation (as plaintiffs would have us do here) against the legitimate purpose sought to be achieved by the legislature. Cf. McGinnis v. Royster, 410 U.S. 263, 93 S. Ct. 1055, 35 L. Ed. 2d 282 (1973). To do so would mean the re-incarnation of a doctrine that the Supreme Court laid to rest in Dandridge v. Williams, 397 U.S. 471, 484, 90 S. Ct. 1153, 25 L. Ed. 2d 491 (1970): the doctrine that the Fourteenth Amendment gave the judiciary leave to strike down state laws which the court found to be "unwise, improvident, or out of harmony with a particular school of thought." Williamson v. Lee Optical Co., 348 U.S. 483, 488, 75 S. Ct. 461, 464, 99 L. Ed. 563 (1955).

"So long as its judgments are rational, and not invidious, the legislature's efforts to tackle the problems of the poor and needy are not subject to a constitutional straitjacket. The very complexity of the problems suggest that there will be more than one constitutionally permissible method of solving them," Jefferson v. Hackney, 406 U.S. 535, 546-547, 92 S. Ct. 1724, 1731, 32 L. Ed. 2d 285 (1972).

Furthermore, it is entirely consistent with the Equal Protection Clause, for a state to "take one step at a time, addressing itself to the phase of the problem which seems most acute to the legislative mind" even if to do so means that other phases of the problem are neglected. Williamson v. Lee Optical, 348 U.S. 483, 489, 75 S. Ct. 461, 465, 99 L. Ed. 563 (1955).

Therefore, the sole question before this court is whether the classification adopted by the legislature bears a rational relationship to a legitimate governmental interest. Frontiero v. Richardson, supra; Dandridge v. Williams, supra; Massachusetts Board of Retirement v. Murgia, supra; Geduldig v. Aiello, 417 U.S. 484, 495, 94 S. Ct. 2485, 41 L. Ed. 2d 256 (1974). "A statutory discrimination will not be set aside if any state of facts reasonably may be conceived to justify it." McGowan v. Maryland, 366 U.S. 420, 426, 81 S. Ct. 1101, 1105, 6 L. Ed. 2d 393 (1961); Dandridge v. Williams, supra, 397 U.S. at 485, 90 S. Ct. 1153.

We need look no further than the legitimate desire on the part of the state to conserve limited public resources in order to find a rational basis for the $80 limit on deductions. See Dandridge v. Williams, supra; Jefferson v. Hackney, supra; cf. Geduldig v. Aiello, supra; Hughes v. Alexandria Scrap, ___ U.S. ___, 96 S. Ct. 2488, 49 L. Ed. 2d 220 (1976). When the original limit of work-related expenses of $60 was raised to $80 in 1974, the additional cost to New York State was over $1 million dollars. The state could rationally have concluded that to eliminate the expense limitation completely or to raise it further, would result in larger expenditures of limited funds.

Plaintiffs have argued that because of the work disincentive built into the program, the state would actually save money by eliminating the expense limitation. This argument, however, is addressed to the wrong forum. "[T]he Constitution does not empower . . . [the] Court[s] to second-guess state officials charged with the difficult responsibility of allocating limited public welfare funds among the myriad of potential recipients." Dandridge v. Williams, supra, 397 U.S. at 487, 90 S.Ct. at 1163.

Due Process

Plaintiffs also challenge the statute on due process grounds, arguing that it creates a presumption that money actually spent on work expenses in excess of $80 per month is available to meet other needs. It is claimed that the presumption is arbitrary, capricious, erroneous and irrebuttable by its terms.

We do not agree that the limitation violates the Due Process Clause. In Richardson v. Belcher, 404 U.S. 78, 81, 92 S. Ct. 254, 257, 30 L. Ed. 2d 231 (1971) the Supreme Court concluded that a classification which "meets the test articulated in Dandridge is perforce consistent with the due process requirement of the Fifth Amendment." Since a non-contractual claim to receive funds from the public treasury is not constitutionally protected, Weinberger v. Salfi, 422 U.S. 749, 772, 95 S. Ct. 2457, 45 L. Ed. 2d 522 (1975), in evaluating a social welfare program, "the Due Process Clause can be thought to interpose a bar only if the statute manifests a patently arbitrary classification, utterly lacking in rational justification." Flemming v. Nestor, 363 U.S. 603, 611, 80 S. Ct. 1367, 1373, 4 L. Ed. 2d 1435 (1960).

The limitation on work related expenses is not patently arbitrary nor utterly lacking in rational justification. The state makes no claim of allowing all deductions for work related expenses. Rather it has balanced the desire to provide incentives to work against use of the state's limited resources. The limit of $80 set by the legislature was apparently based to some extent upon a memorandum of the Deputy Commissioner of the New York State Department of Social Services which estimated average monthly work expenses at $75, although indicating that working persons often incurred more than $100 in work-related expenses. Whether or not we agree with the final decision of the legislature to place the maximum limitation for work expenses at $80, we must remain cognizant of the fact that "every line drawn by a legislature leaves some out that might well have been included. That exercise of discretion, however, is a legislative, not a judicial function." Village of Belle Terre v. Boraas, 416 U.S. 1, 8, 94 S. Ct. 1536, 1541, 39 L. Ed. 2d 797 (1973); see Louisville Gas Co. v. Coleman, 277 U.S. 32, 41, 48 S. Ct. 423, 72 L. Ed. 770 (Hughes, J., dissenting).

We think it appropriate to re-emphasize what the Supreme Court said in Dandridge v. Williams:

We do not decide today that the . . regulation is wise, that it best fulfills the relevant social and economic objectives that . . . [the state] might ideally espouse, or that a more just and humane system could not be devised. Conflicting claims of morality and intelligence are raised by opponents and proponents of almost every measure, certainly including the one before us. But the intractable economic, social, and even philosophical problems presented by public welfare assistance programs are not the business of this Court. The Constitution may impose certain procedural safeguards upon systems of welfare administration, [citation omitted]. But the Constitution does not empower this Court to second-guess state officials charged with the difficult responsibility of allocating limited public welfare funds among the myriad of potential recipients. [citations omitted]

397 U.S. at 487, 90 S.Ct. at 1162. Cf. Lavine v. Milne, 424 U.S. 577, 96 S. Ct. 1010, 47 L. Ed. 2d 249 (1976).

Since we conclude that New York State Social Services Law § 131-i and 18 N.Y.C. R.R. § 352.19(c) do not violate the Constitution, plaintiffs' motion for a permanent injunction is denied and the complaint is dismissed. The Clerk of the court is directed to enter judgment accordingly.

WEINSTEIN, District Judge (dissenting).

Our society respects honest work for pay. The work ethic is fundamental in our culture. Empirical studies demonstrate that able-bodied persons in all income brackets prefer work to relief. Government programs providing aid to some of those in need have been designed on the assumption that each recipient household will support itself to the extent possible. A law that strips people of their dignity by encouraging them to stop working in order to qualify for enough financial assistance to bring family income up to bare subsistence levels is inconsistent with fundamental policy. To put it plainly, it is irrational. Joined with other discriminations against poor, two-parent families, all contributing to the destruction of family life, the impact of the welfare provision in issue verges on the absurd.

At the heart of this case is the acknowledged fact that the working families in the plaintiff class have been forced to live on less than the statutorily defined minimum subsistence level of income which New York guarantees to its unemployed welfare recipients simply because their monthly job-related expenses exceed $80.00. The challenged statute—New York Social Services Law § 131-i—by limiting work expense deductions to $80 per month, requires plaintiffs to report to the administering welfare agencies net monthly incomes higher than their actual disposable incomes. Plaintiffs are consequently denied benefits for which they would otherwise be eligible under the provisions of New York's Home Relief program. New York Social Services Law §§ 157 et seq.

The work expense ceiling, enacted subsequent to the institution of the Home Relief program, effects an irrational, arbitrary discrimination against the working poor violative of the Equal Protection Clause of the Fourteenth Amendment. This legislative afterthought simply grafts a set of gratuitous inequities onto a pre-existing welfare program, exacerbating the system's proven tendencies to discourage work initiative and to destroy family life.

We recognize that states do have wide discretion in the field of welfare legislation and that the reviewing authority of the federal courts is limited. See, e.g., Dandridge v. Williams, 397 U.S. 471, 90 S. Ct. 1153, 25 L. Ed. 2d 491 (1970). Federal courts, mindful of these restrictions, have been reluctant to enforce the standard of "minimum rationality" against state legislatures. See, e.g., Tribe, Structural Due Process, 10 Harv.Civil Rights L.Rev. 269, 272 (1975); Linde, Due Process of Lawmaking, 55 Neb. L.Rev. 197 (1976). Yet, unless that doctrine is dead in all but such strict scrutiny cases as those involving race, this case warrants its application. Financial savings alone cannot justify an unprincipled classification, serving no other legitimate state interest. Accordingly, for the reasons set forth below, I dissent.

I. Challenged State Provisions in Context

Plaintiffs' complaint is directed at one small feature of New York's Home Relief program, Social Services Law § 131-i, setting a limit on the deductibility of work expenses in calculating the income level of applicants.

Home Relief is New York State's residual public assistance program. It is intended to provide a subsistence level of income to those of the State's needy who are ineligible to participate in the categorical welfare programs such as Aid to Families with Dependent Children (AFDC) and Supplemental Security Income (SSI). Id. § 158.

Recipients of Home Relief are characteristically needy adults and children in two-parent families who are neither aged, blind, nor permanently and totally disabled. Comparably needy children in one-parent families are eligible for AFDC. Id. § 348 A & B. Needy aged, blind, or otherwise permanently and totally disabled persons are eligible for SSI pursuant to 42 U.S.C. §§ 1381 et seq. and New York Social Services Law §§ 209 et seq.

The amount of Home Relief paid to a recipient family equals the difference between its net monthly income from other sources and its prescribed level of need. "Need" is determined by adding the actual cost of shelter (within certain maximums related to family size) to an amount representing the uniform "statewide standard of need" for a family of specific size. Social Services Law § 131-a 2 sets forth the various standards which are applicable to both Home Relief and AFDC. The statewide standard of need represents the State legislature's determination of the subsistence level for all items, except shelter, needed by humans such as food, clothing, electricity, telephone, transportation, household goods, toiletries, school supplies, replacement furniture, books and other reading matter, and special allowances when appropriate in specific cases—e. g., an allowance for meals taken in a restaurant when the aid recipient lacks cooking facilities or is physically unable to cook. See, generally, Rosado v. Wyman, 304 F. Supp. 1356, 1365 (E.D.N.Y. 1969), vacated, 414 F.2d 170 (2d Cir. 1969), reversed in part, 397 U.S. 397, 90 S. Ct. 1207, 25 L. Ed. 2d 442 (1970).

Prior to June 14, 1972, applicable Social Service regulations, 18 NYCRR § 352.19 et seq., authorized the deduction in full of all necessary work-related expenses from gross income to determine both the financial eligibility of working applicants for supplementary Home Relief and the amount of such relief to be granted. Deductible expenses included so-called "personal work expenses", such as Federal, State, and local income taxes, Social Security and other payroll deductions, meals taken at work, and transportation, 18 NYCRR § 352.19(2); "non-personal work expenses", such as the cost of union dues, tools, uniforms, permits, and licenses, 18 NYCRR § 352.19(1); and a "special work expense" deduction of $20 a month to meet other additional or unforeseeable costs necessary and incident to employment, 18 NYCRR § 352.19(3). Expenses for all of these items were allowed in their actual amount except that, then as now, the allowances for lunch or dinner taken during the work day were fixed at maximums of $1.00 and $1.45 respectively. Assuming the existence of all other eligibility factors, a deficit between recognizable need and net available income, thus computed, rendered the family eligible for Home Relief. The amount of the deficit determined the amount of the grant. (Participants in a certified training program were and presently remain eligible for an additional $30 exemption per month, Social Services Law § 159-b.)

In 1971, the New York Legislature amended the Social Services Law by placing a $60 per month ceiling on the deduction of work-related expenses. Laws 1971, Chap. 746, § 2, eff. July 1, 1971. This ceiling was raised to the present $80 per month in 1974. Laws, 1974, Chap. 1059, § 1, eff. September 1, 1974. No changes have been made since then although we take judicial notice of the fact that the cost of living and, presumably, the cost of working having risen substantially in the interim. Such costs as union dues, transportation, tools and uniforms are likely to have risen over the past two years.

The limitation is presently codified as Social Services Law § 131-i, reading as follows:

§ 131-i. Expenses necessary and incident to employment; maximum allowances from earnings

Whenever any applicant for or recipient of public assistance is employed, provision may be made for allowing appropriate amounts from his earnings for expenses necessary and incident to his employment, but in no case shall such allowance exceed eighty dollars per month unless required by federal law or regulation. The allowance provided herein shall not apply with respect to income earned by recipients of home relief in public work projects.

The implementing regulation, 18 NYCRR § 352.19(c), was amended to reflect the new $80.00 monthly ceiling on October 4, 1974, effective September 1, 1974. It now reads:

(c) Work expenses in Home Relief cases. Notwithstanding the provisions contained in subdivisions (a) and (b) of this section, an allowance for all personal, non-personal and special work expenses necessary and incidental to employment shall in no case exceed $80 per month for an applicant for or recipient of home relief. The allowance provided herein shall not apply with respect to income earned by recipients of home relief in public work projects.

At present an applicant for or recipient of Home Relief with monthly work-related expenses in excess of $80 has only $80 deducted from earnings to compute net recognizable income. If this net income figure exceeds the statewide standard of need for the family size, the family is ruled ineligible for Home Relief. It is then denied the automatic full Medicaid coverage bestowed on all recipients of Home Relief. Social Services Law § 366.

II. Impact on Plaintiffs

The work expense ceiling has had a harsh impact on the named plaintiffs. Levan Roundtree, a truck driver, is the sole source of support for his family of six. His gross monthly salary is $714.14; his monthly work-related expenses, consisting of Federal, State, and City income taxes, Social Security tax, New York State Disability Insurance, union dues, transportation, lunch, and work clothing amount to $156.66. Because only $80 in expenses could be deducted, Roundtree's application for supplemental Home Relief was denied. But, when monthly mortgage payments, property taxes, and carrying charges on their home are deducted, the Roundtrees are left with $250.48 in monthly available income, only 68% of the $368.00 welfare subsistence level for a family of six. The Roundtrees are also, unlike those receiving welfare payments, ineligible for Medicaid coverage (Social Services Law § 366 1(a)(1), 2(a)(8)(ii))— which may have a value of thousands of dollars a year.

Similarly, the Pelligrinos, a family of nine, must subsist on a monthly income, after rent, of $485.34—$32.66 below the statewide standard—because Mr. Pelligrino, a mason's helper for a construction company, incurs $112.67 in work expenses ($69.33 in involuntary payroll deductions and $21.67 for transportation and lunch). The Folsoms, a family of seven, experience a $23.56 subsistence deficit because Mr. Folsom, a janitor, incurs $103.56 in monthly work expenses. Included are $64.06 in involuntary payroll deductions, despite the fact that because of his low income and number of dependents Mr. Folsom pays no federal income tax.

Data available to the court suggests that the work-related expenses of these families are typical. A 1972 study of the New York welfare system, using 1971 tax schedules and typical 1970 union dues, shows that the work expenses for the head of a four-person household working 40 hours a week at $2 per hour were even then as high as $800 annually, or $67 per month. See B. Bernstein, A. N. Shkuda, M. Burns, "Income-Tested Social Benefits in New York: Adequacy, Incentive, and Equity," in Studies in Public Welfare, Paper No. 8 prepared for the Subcommittee on Fiscal Policy Joint Economic Committee 19 (Washington, D.C.: Government Printing Office, July 8, 1973). This figure, which increases as income rises, does not reflect many expenses recognized by 18 NYCRR § 352.19, such as tools, materials, fees, and license payments.

A study of 560 families then receiving Home Relief in 1971 showed that work-related expenses averaged $75 per month and that "working persons often incur $100 or more per month in necessary work-related expenses." Memorandum from Felix Infausto, Deputy Commissioner, New York State Department of Social Services, to the State Legislature in Support of Assembly, Bill No. 11310, Appendix A to Plaintiffs' Memorandum of Law in Support of Their Motion for a Preliminary Injunction. Given inflation and the increases since 1971 in both mass transit fares and State, City and Social Security taxes, it is doubtful that the present $80 figure accurately reflects the work-related expenses of the average New York worker otherwise eligible for Home Relief.

Much of what follows elaborates the extent to which present laws disadvantage plaintiffs and their class relative to those receiving higher benefits. Nothing said below should be taken as suggesting that welfare recipients lead lives of idle affluence. We take judicial notice, based on many civil and criminal cases tried in this court involving poor people, that only a small percentage of those on welfare receive the high benefits that would result from the addition of all theoretically available programs. Recipients of welfare assistance are generally economically and otherwise deprived.

III. Equal Protection

The imposition of a ceiling on monthly work-related expenses has created new inequities in the treatment of Home Relief applicants and recipients. In effect, the statute now distinguishes (1) claimants who are employed and incur work expenses of more than $80 monthly from (2) those who are employed and incur monthly work expenses of $80 or less and those who are unemployed. Persons in the latter class are guaranteed a disposable monthly income—i. e., income after deduction of work-related expenses—equal to the statutory subsistence standard. Persons in the former class—though in all relevant respects identically situated—receive diminished benefits or no benefits at all and are thus left with disposable monthly incomes below the statutory standard.

Hypothetical examples readily illustrate the mechanics of the discrimination. Assume that four two-parent, five-member families are each paying $200 in rent. The recognized monthly "need" for each family would be $518. If no one works in family A, it will receive $518 in Home Relief. If one person works in family B, earning a gross monthly salary of $550 and incurring $80 in work expenses, B will receive a $48 Home Relief grant, giving it $518 in monthly disposable income. If, however, one person in family C incurs $95 in monthly work expenses while earning the identical $550 in gross income, the $80 work expense ceiling will result in C's receiving the same $48 supplement and, hence, $15 less in monthly disposable income. Finally, if one person in family D earns $600 gross, but incurs $110 in monthly work expenses, D will receive no Home Relief since the $80 ceiling will cause its monthly net disposable income level to be recognized as $520 per month, or $2 above the statutory standard. In reality, though, D's disposable income is $490, a full $28 below the standard. Ironically, family D, because it is ineligible for Home Relief, is the only family ineligible for full Medicaid benefits—even though in real terms it is the neediest of the four. Social Services Law § 366 1(a)(1), 2(a)(8)(ii).

It is axiomatic that the Equal Protection Clause of the Fourteenth Amendment requires some rational justification for legislative discrimination. Under traditional equal protection principles, a state retains broad discretion to classify—particularly in the area of economics and social welfare— provided its classification has "a reasonable basis." See, e. g., Graham v. Richardson, 403 U.S. 365, 371, 91 S. Ct. 1848, 1852, 29 L. Ed. 2d 534 (1971); Dandridge v. Williams, 397 U.S. 471, 485, 90 S. Ct. 1153, 1161-62, 25 L. Ed. 2d 491 (1970); McGowan v. Maryland, 366 U.S. 420, 425-27, 81 S. Ct. 1101, 1106, 6 L. Ed. 2d 393 (1961); Williamson v. Lee Optical Co., 348 U.S. 483, 489, 75 S. Ct. 461, 465, 99 L. Ed. 563 (1955). "[A]s in all equal protection cases, . . . the crucial question is whether there is an appropriate governmental interest suitably furthered by the differential treatment." Police Department of Chicago v. Mosley, 408 U.S. 92, 95, 92 S. Ct. 2286, 2290, 33 L. Ed. 2d 212 (1972). "[A] State cannot achieve its objectives by totally arbitrary means; the criterion for differing treatment must bear some relevance to the object of the legislation." Bullock v. Carter, 405 U.S. 134, 145, 92 S. Ct. 849, 857, 31 L. Ed. 2d 92 (1972).

Budgetary considerations, standing alone, do not afford a constitutionally sufficient justification for legislative discrimination which disadvantages a single group. See Graham v. Richardson, 403 U.S. 365, 372-76, 91 S. Ct. 1848, 1852-53, 29 L. Ed. 2d 534 (1971); Shapiro v. Thompson, 394 U.S. 618, 631, 89 S. Ct. 1322, 1330, 22 L. Ed. 2d 600 (1969); see also Bullock v. Carter, 405 U.S. 134, 149, 92 S. Ct. 849, 858, 31 L. Ed. 2d 92 (1972); Rinaldi v. Yeager, 384 U.S. 305, 86 S. Ct. 1497, 16 L. Ed. 2d 577 (1966). In Shapiro, the Court specifically noted that the saving of welfare costs cannot justify an invidious classification. It said:

We recognize that a State has a valid interest in preserving the fiscal integrity of its programs. It may legitimately attempt to limit its expenditures, whether for public assistance, public education, or any other program. But a State may not accomplish such a purpose by invidious distinctions between classes of its citizens. It could not, for example, reduce expenditures for education by barring indigent children from its schools. Similarly, in the cases before us, appellants must do more than show that denying welfare benefits to new residents saves money. The saving of welfare costs cannot justify an otherwise invidious classification.

394 U.S. at 631, 89 S.Ct. at 1330.

Although in recent welfare cases the Court has spoken sympathetically of state needs to preserve limited fiscal resources, it has consistently recognized the constitutional requirement that cost savings be effectuated on a principled basis. Thus, in Dandridge v. Williams, supra, applying a limited rational relationship test, it looked to policy objectives beyond fiscal considerations in upholding Maryland's challenged AFDC benefit ceiling—specifically, the State's interests "in encouraging employment and in avoiding discrimination between welfare families and the families of the working poor." 397 U.S. at 486, 90 S.Ct. at 1162. As indicated below, both these objectives are defeated by the classification challenged here. Cf. Hughes v. Alexandria Scrap, ___ U.S. ___, 96 S. Ct. 2488, 49 L. Ed. 2d 220 (1976) (Maryland, in preserving limited funds available for program paying bounties to scrap-processors for destroying abandoned automobiles may constitutionally impose additional documentation requirements on out-of-state processors only since that discrimination is rationally related to the basic statutory purpose of cleaning Maryland's landscape of abandoned automobiles).

Similarly, in Jefferson v. Hackney, 406 U.S. 535, 549, 92 S. Ct. 1724, 1733, 32 L. Ed. 2d 285 (1972), Texas' decision to reduce its AFDC standard of need by a greater percentage than it reduced the standard used in its other federally aided categorical assistance programs was upheld as furthering its legitimate interest in maximizing the effectiveness of its fixed pool of welfare funds. The Court observed that:

The State may have concluded that the aged and infirm are the least able of the categorical grant recipients to bear the hardships of an inadequate standard of living. . . . [and] that the young are more adaptable than the sick and elderly . . . .

The members of the plaintiff class share no comparable attributes differentiating them from the favored Home Relief recipients. In this sense, the classification here is as arbitrary as a reduction of benefits for every third recipient on the welfare rolls.

It should be emphasized that New York singled out plaintiffs' sub-class to bear the burden of reduced benefits after its initial Home Relief program had treated all recipients equally. This case is strikingly different from those in which state benefit programs have been upheld against charges of under-inclusiveness in their original design. Cf. Geduldig v. Aiello, 417 U.S. 484, 94 S. Ct. 2485, 41 L. Ed. 2d 256 (1974). The proposition that "reform may take one step at a time, addressing itself to the phase of the problem which seems most acute to the legislative mind", Williamson v. Lee Optical Co., 348 U.S. 483, 489, 75 S. Ct. 461, 465, 99 L. Ed. 563 (1955), is irrelevant to the question presented here. See also, Katzenbach v. Morgan, 384 U.S. 641, 656-57, 86 S. Ct. 1717, 1726-27, 16 L. Ed. 2d 828 (1966).

New York may reduce benefit levels across the board on an equal percentage basis. See New York Laws 1971, Ch. 133, § 2, eff. May 1, 1971 (reducing benefits to 90% of the statewide standard of need); Rosado v. Wyman, 397 U.S. 397, 90 S. Ct. 1207, 25 L. Ed. 2d 442 (1970). But the validity of its decision to place the entire burden of cost savings on a narrowly defined subgroup turns on whether that classification reflects some rationale beyond the mere reduction of program costs. Shapiro v. Thompson, 394 U.S. 618, 89 S. Ct. 1322, 22 L. Ed. 2d 600 (1969); Dandridge v. Williams, 397 U.S. 471, 90 S. Ct. 1153, 25 L. Ed. 2d 491 (1970).

Our inquiry is, admittedly, limited, for as the Court noted in Dandridge, "[a] statutory discrimination will not be set aside if any state of facts reasonable may be conceived to justify it." 397 U.S. at 485, 90 S.Ct. at 1161. But see Green v. Waterford Board of Education, 473 F.2d 629, 633 (2d Cir. 1973) (noting trend in recent Supreme Court decisions toward more rigorous application of rational relationship test). Any assessment of the impact of this statute, taking into account the practical realities of life on welfare, the overall structure of the various programs, and the findings of experts, demonstrates that the work expense ceiling is at best irrelevant to, and at worst destructive of, the fundamental objectives of welfare policy.

IV. Policy Objectives

We may posit several possible objectives supporting any given change in the welfare law: (A) reduction of poverty; (B) elimination of horizontal inequities — particularly as between the working poor and unemployed recipients; (C) promotion of administrative efficiency; (D) encouraging recipients to increase income through employment; (E) strengthening the family ties of program participants; and (F) reduction of program costs. See, e. g., Dandridge v. Williams, 397 U.S. 471, 486, 90 S. Ct. 1153, 1162, 25 L. Ed. 2d 491 (1970); Shapiro v. Thompson, 394 U.S. 618, 633-38, 89 S. Ct. 1322, 1331-33, 22 L. Ed. 2d 600 (1970). The challenged provision contributes nothing to the advancement of such goals.

A, B, and C. Reduction of Poverty, Inequalities, and Administrative Inefficiency.

The work expense ceiling does nothing to ease the pressure of poverty. As indicated below, it increases the inequities within the Home Relief program. There is certainly no reason to assume that those whose work-related expenses happen to exceed $80 are better suited than other Home Relief recipients to living on monthly incomes below the subsistence standard. Similarly, there is no gain in administrative efficiency since the $80 figure is only a ceiling and the administering social services worker must still compute the actual amount of work expenses in each case. 18 NYCRR § 352.19 et seq. The same procedures would be followed if actual work expenses were fully deductible.

D. Work Incentives.

(1) Home Relief Program

The Home Relief program is a guaranteed income plan. Each reported dollar earned below the statutory standard of need results in the loss of a dollar in benefits. There is no economic incentive for an unemployed Home Relief recipient to accept paid employment since it affords no net gain in retained income. Thus the "benefit-loss rate" or "tax" on net earned income would be 100%.

For the members of the plaintiff class, acceptance of paid employment results in a net decrease in disposable income because the loss of benefits is not adjusted to reflect all the income that is lost in the form of work expenses. The benefit-loss rate for plaintiffs is thus greater than 100%. Some employed Home Relief recipients have a strong incentive to reduce earnings or avoid increases in salary since resulting increases in income-related work expenses — such as income and Social Security taxes — may generate net decreases in disposable income. Losses from related welfare programs — typically medical assistance, food stamps, and school lunches — can substantially enhance the disincentives to work.

Those generalizations are supported by studies of the operation of the New York welfare system. See, e. g., B. Bernstein, N. Shkuda, M. Burns, "Income-Tested Social Benefits in New York: Adequacy, Incentive, and Equity," in Studies in Public Welfare, Paper No. 8 prepared for the Subcommittee on Fiscal Policy, Joint Economic Committee (Washington, D.C.: Government Printing Office, July 8, 1973). The Bernstein-Shkuda-Burns study focuses on the "notches" in the programs, i. e., the points at which small increases in work income produce greater reductions in benefit payments because of failure to account for taxes and work expenses.

There are serious notches for the intact family on home relief. . . . When the intact family's gross income increases from $3,000 to $4,000, equal to a $691 gain in disposable income when taxes and work expenses are taken into account, welfare benefits are reduced by $922. A further income increase to $5,000, a $773 gain in disposable income, results in the loss of a $632 welfare grant.

Id. at 4. Elaborating further, the authors note:

When moving into employment (earning $4,160 annually) after training the Home Relief family of four has a net gain of only $279 over the amount it would receive from welfare. Further, the family's disposable income does not increase as earnings rise from $4,160 to $5,000. In fact, as earnings increase from $4,160 to $5,000, disposable income declines from $4,191 to $4,002. The disincentive for the family head who has not gone through a training program and therefore does not benefit from the $30 disregard is even more severe; annual earnings of $5,160 would leave him with a disposable income of $3,831, or somewhat less than he would receive if wholly dependent on welfare. Further, earnings of $4,500 or $4,700 also yield less disposable income than complete dependence on welfare. At the $5,000 level, the Home Relief family, whether benefiting from the $30 disregard or not, has only $90 more in annual disposable income than by relying solely on assistance. . . .

Id. at 28.

The relevant data for the intact family make it strikingly clear that there is practically no advantage . . . to the intact four-person family in increasing earnings from zero up to $7,000 gross income in relation to the basic social programs including medical assistance. At various points on the income scale it is worse off. Such a family ends up with only $180 more per year if it earns $3,000 that if it has no earnings. Because of the notch in the public assistance program at $4,000, the family which moves from zero earnings to gross earnings of $4,000 is slightly worse off and if it moves from $3,000 to $4,000 it loses about $230 in the process. The notch in the medicaid program at $5,000 due to rules regarding outpatient care results in a net loss of about $200 for the family which raises its earnings from zero to $5,000. At best, if the family is living in subsidized housing, it will be $144 better off with earnings of $5,000 than with no earnings at all.

Even at $6,000 of earnings the family with "average" medical costs of $1,200 will be no better off than at zero earnings and at $7,000 it will only have about $200 more a year. Only if the family's medical expenses are substantially less than average does it make any substantial gain by increasing its income from $6,000 to $7,000.

As earnings increase above $8,000, the family does, of course, obtain a larger net benefit from each $1,000 increase in income, and the difference in the level of living achieved by those with earnings of $8,000 or more and those relying exclusively on social programs becomes significant. But for those who, because of limited education and skills, or because of discrimination, or other reasons, are unable to earn more than $6,000 to $7,000 per year, the present set of eligibility criteria for benefits from the major social programs does not provide any incentive to increase income. In fact, the result is quite the contrary.

Id. at 140-42 (emphasis added).

The authors conclude:

In summary, one is led to conclude that the benefits from the package of social programs available to New Yorkers under current rules and regulations constitute a disincentive to the acceptance of relatively low-paying or even not so low-paying jobs. Further, they are a disincentive to increasing family income by working longer hours or by a second adult in the family entering the labor market unless the family can look forward to achieving an income of more than $8,000 a year. Indeed, the notch problems present in each of the programs tend to accumulate and to reinforce the disincentive effect.

Id. at 146 (emphasis added).

The figures used in this study are dated. Since 1972, benefit levels in all programs have increased, as have wage levels, City, State, and Social Security tax rates, mass transit fares, and other work-related expenses. These changes have all inured to the relative detriment of the working poor family vis à vis the unemployed welfare recipient. Wage increases are still matched by benefit loss; but increases in expenses over the ceiling result in even greater losses of disposable income. See, Senator William T. Smith, "Public Welfare — the ``Impossible' Dream," Address presented at the County Officers Association of the State of New York 1975 Annual Fall Seminar, September 22, 1975 (citing inequities and disincentives, using more recent figures).

(2) Empirical Studies

The work response of Home Relief recipients to the disincentives described apparently has not been systematically studied. However, studies of other programs in which benefit loss cancels out added earned income have found sharp reductions in work. A 1960 study of Social Security recipients demonstrated that the imposition of a reduction in benefits for men earning more than a specified amount caused men reaching age 65 to drop out of the labor force "at a precipitous rate." See V. Burke and A. Townsend, "Public Welfare and Work Incentives: Theory and Practice" in Studies in Public Welfare, Paper No. 14 prepared for the Subcommittee on Fiscal Policy, Joint Economic Committee 31 (Washington, D.C.: Government Printing Office, April 15, 1974). At age 72, when the penalty is removed, labor force participation increased. In the absence of other compelling reasons, the study concluded that the elimination of the tax on earnings caused more people to work. Id.

A separate study of Social Security found that in 1965, when the amount of earnings exempt from benefit loss was raised from $1200 to $1500, about 10% of both male and female Social Security retirees increased their earnings from just below $1200 to just below $1500. The author attributed this $300 earnings jump to the removal of the tax on this sum. W. Vroman, Older Worker Earnings and the 1965 Social Security Amendments (Washington, D.C.: U. S. Department of Health, Education and Welfare, Social Security Administration, Office of Research and Statistics, Report No. 38, 1971) summarized in Burke and Townsend, supra at 31-32.

Similar findings were made in a recent study of Unemployment Insurance in Wisconsin. See, R. Munts, "Partial Benefits in Unemployment Insurance: Their Effect on Work Incentive," 4 Journal of Human Resources 160-76 (Spring 1976) summarized in Burke and Townsend, supra at 32. Under that program full benefits were paid so long as the claimant earned less than one half the benefit amount. At the half way point benefits were reduced to one half until earnings equalled the full benefit amount, at which point benefits ceased. Thus combined earnings and benefits were maximized at earning levels just below half the weekly benefit and just below the full weekly benefit. Analysis of the distribution of earnings of applicants for partial benefits in Wisconsin found that earnings were bunched at these points. "This suggests that these U.I. claimants could have worked more but were reluctant to have additional earnings serve only to offset their U.I. benefits dollar for dollar." Burke and Townsend, supra at 32.

If such earning reductions occurred in all welfare programs, then the disincentives built into the Home Relief program could be constitutionally justified as a necessary by-product of a legitimate policy of income maintenance. But economic theory recognizes, and behavioral research corroborates, that welfare programs can be designed without diminishing the natural inclination of participants to increase income beyond mere subsistence through paid employment.

Economic analysis posits that the imposition of a benefit loss or tax rate on earnings substantially less than 100% (say 50%) may have either of two opposite effects on the behavior of a welfare recipient: (1) a substitution effect inducing substitution of leisure for work and (2) an income or marginal dollar effect, reducing the individual's ability to afford leisure and encouraging greater work efforts. For discussions of the income and substitution effects in economic theory, see, e. g., Burke and Townsend, supra at 9-10; I. Garfinkel, "Income Transfer Programs and Work Effort: A Review" in Studies in Public Welfare, Paper No. 13 prepared for the Subcommittee on Fiscal Policy, Joint Economic Committee (Washington, D.C.: Government Printing Office: Feb. 18, 1974).

A benefit loss rate of 100% or more — as in Home Relief — wipes out the income effect, since no marginal dollars can be earned by increased work effort. Work reduction is the only rational response. But for benefit loss rates below 100% economic theory is consistent with behavior changes in either direction. Which force actually predominates is an empirical question which can be answered only by analysis of actual work patterns of recipients and potential recipients. See, e. g., V. Burke and A. Townsend, "Public Welfare and Work Incentives: Theory and Practice" in Studies in Public Welfare, Paper No. 14 prepared for the Subcommittee on Fiscal Policy, Joint Economic Committee 8 (Washington, D. C., Government Printing Office, April 15, 1974).

The most significant available study of the effects of variable minimum guarantees and benefit-loss rates on work effort is the New Jersey Income Maintenance Experiment, which ran for three years beginning in August, 1968. See, Summary Report: New Jersey Graduated Work Incentive Experiment (U.S. Department of Health, Education, and Welfare, December, 1973) for a detailed report of the experiment; see also Burke and Townsend, supra at 33-35; I. Garfinkel, "Income Transfer Programs and Work Effort: A Review" in Studies in Public Welfare, Paper No. 13, Subcommittee on Fiscal Policy, Joint Economic Committee (Washington, D.C.: U. S. Government Printing Office, Feb. 18, 1974); S. R. Wright and J. D. Wright, "Income Maintenance and Work Behavior," 6 Social Policy 24 (1975). The experiment divided male-headed families whose normal income fell below 150% of the poverty line into eight different welfare programs and a control group that received no benefits. The programs were designed with varying minimum guarantee levels and benefit-loss rates ranging from 30 to 70 percent. The researchers found extremely small differences in average hours worked between experimental and control group husbands and no significant differences in the percentage of husbands who did not work at all during any of the three years of the experiment:

The most striking features of the results for husbands . . . are that all of the differentials are quite small in both absolute and relative terms — none exceed 10 percent of the control means and most are less than five percent — and all are statistically insignificant . .. There are no findings here to indicate a significant reduction in labor supply resulting from the experimental payments. Moreover, many of the differentials, including all of those for blacks, are positive, indicating greater labor supply among husbands in the treatment group than in the control group. Finally, it is worth noting that the means for both groups indicate that the vast majority (approximately 95 percent) of the husbands were labor force participants who when employed worked close to full-time (37 to 40 hours per week).

Summary Report, supra, at 19-20.

The absence of significant reductions may be attributed, at least in part, to the fact that under the negative income tax plans studied, earnings result in net increases in income. Id. at V, 5. The evidence thus lends substantial empirical support to the proposition that at tax rates below 100%, the theoretical work-reducing "substitution" effect is counterbalanced by a work-increasing "income" effect. This is particularly true for male heads of households — representing the group characteristically receiving Home Relief. The experiment found more significant levels of work reduction among experimental group wives (although in absolute terms these reductions were also small). Id. at 32-35. The authors conclude:

To the extent that the results of the experiment can be generalized to the national low-income population, they indicate that a national program of income-conditioned transfers, at the benefit levels considered here, would have only relatively small effects on the labor supply of male family heads.

Id. at 32.

Additional evidence supports the New Jersey findings. At least two studies have attributed increases in the employment rates of AFDC recipients to 1967 amendments to the Social Security Act (42 U.S.C. §§ 602(a)(7), 602(a)(8)(A)) allowing recipients who have received benefits within four months prior to employment to deduct, as a work incentive, the first $30.00 of monthly gross earned income plus one-third of the remainder in determining the assistance grant. See G. Appel, Effects of a Financial Incentive on AFDC Employment: Michigan's Experience Between July 1969 and July 1970 (Minneapolis' Institute for Interdisciplinary Studies, March 1972); V. Smith, Welfare Work Incentives, Studies in Welfare Policy, No. 2 (Lansing: Michigan Department of Social Services, 1974), summarized in W. Bell and D. Bushe, Neglecting the Many, Helping the Few: the Impact of the 1967 Work Incentives at 15-21 (Center for Studies in Income Maintenance Policy, New York University School of Social Work, 1975); but see M. Rein, Work or Welfare? Factors in the Choice of AFDC Mothers, Ch. 4 (1974).

Cross-sectional studies using econometric analysis of labor market response to variable wage rates and levels of unearned income to project work response to different income transfer programs have concluded that, at tax rates below 100%, changes in work effort would be small. See Garfinkel, supra at 16-20 for a review of this literature.

Finally, 1971 Census data on employment demonstrate high levels of positive attachment to the labor force for both male and female heads of poor families. M. Barth, G. Carcagno, J. Palmer, Toward an Effective Income Support System; Problems, Prospects and Choices 61 (1974).

These results are consistent with general assumptions about the reluctance of most people to give up the sense of independence and pride that comes with a paying job. Recent attitude surveys find positive work orientations among the poor. See L. Goodwin, Do the Poor Want to Work? (Washington, D.C.: Brookings Institution 1972). Goodwin surveyed more than 4,000 persons through a questionnaire designed to measure a variety of attitudes bearing on work orientation. The survey included long-term AFDC mothers and their teenage sons; short-term AFDC mothers and teenage sons; upwardly mobile black working families in "outer city" areas who had escaped the ghetto through employment; white families in the same outer city census tracts; and welfare recipients enrolled in a national work training program called the Work Incentive Program.

Briefly, in comparing welfare and non-welfare respondents, Goodwin found no statistically significant differences in the work orientation of mothers (id. ch. 3), teenage sons (id. ch. 4) and fathers (id. ch. 5). Significantly, these findings were not qualified by length of time spent on welfare. Goodwin notes:

Teenage males who have spent virtually their entire lives on welfare have certain positive orientations toward work. Having no working parent in the home— neither mother nor father—has made the sons' identification with work no weaker than that of sons from families with working fathers . . .. Welfare youths from fatherless homes show a strong work ethic, a willingness to take training, and an interest in working even if it is not a financial necessity. . . . The welfare experience has not destroyed the sons' positive orientations toward work.

Id. at 68.

These findings are summarized as follows:

Evidence from this study unambiguously supports the following conclusions: poor people—males and females, blacks and whites, youths and adults—identify their self-esteem with work as strongly as do the nonpoor. They express as much willingness to take job training if unable to earn a living and to work even if they were to have an adequate income. They have, moreover, as high life aspirations as do the nonpoor and want the same things, among them a good education and a nice place to live. This study reveals no differences between poor and nonpoor when it comes to life goals and wanting to work.

Id. at 112. Numerous other studies have found additional empirical evidence of positive work orientations among the poor. See generally sources cited in Goodwin at notes 15-20, Ch. I.

The upshot of these studies is that the poor want to work and do work given the most marginal incentive. The financial penalties for earned income built into the Home Relief program are irrational and unjust. They put pressure on those who desire the status of workers to subsist instead solely on welfare. The challenged discrimination thus in no way serves, and may be antithetical to, legitimate employment policies.

E. Family Stability

The $80 work expense ceiling exacerbates the welfare system's inherent discrimination against intact families. Single-parent AFDC families, for example, are entitled to full work expense deductions. 42 U.S.C. § 602(a)(7); see Shea v. Vialpondo, 416 U.S. 251, 94 S. Ct. 1746, 40 L. Ed. 2d 120 (1974). As noted above, 1967 Social Security amendments provided $30 and one-third monthly income disregards for AFDC mothers who have received benefits within four months prior to employment. 42 U.S.C. §§ 602(a)(7), 602(a)(8); 45 C.F.R. 233.20(a)(11)(ii)(b). As a result, female-headed single-parent families fare much better under AFDC than do intact families on Home Relief and intact families ineligible for any assistance. Focusing on these inequities the authors of the Bernstein-Shkuda-Burns study observe:

One fact comes through clearly: Even though New York, unlike some States, has a general assistance program for the intact family, such a family, on the whole, fares worse than the female-headed family; this is true even though the medicaid benefits tend to moderate the inferior public assistance benefit available to the intact family, as do the benefits available from subsidized housing and day care. But both the intact family and the "regular" female-headed household obtain substantially smaller benefits than the female-headed family with the $30 and one-third disregard. The effect of this disregard and the benefits available to a welfare family from food stamps, school lunches, and medicaid are such that with gross earnings of $5,000, the intact family has a disposable income plus benefits of $5,363, the regular female-headed family has $5,892, and the "$30 and one-third" female-headed family has $7,246. The relative benefits are a bit higher for the intact family than for the regular female-headed family at the $6,000 to $7,000 levels because its somewhat higher medical costs are partially covered. But while both types of families with gross earnings of $7,000 have a disposable income plus benefits (other than housing and day care) of $5,780 and $5,675 respectively, the "$30 and one-third" female-headed family has $7,913.

. . . . .

It is worth pointing out that the female-headed family entitled to the $30 and one-third disregard is in a privileged position not only with respect to other types of families on public assistance but in relation to families not eligible for public assistance. Thus, a female-headed family in which the mother is earning $7,000 per year obtains a disposable income plus benefits from public assistance, food stamps, school lunches, and Medicaid of about $7,900 per year. The non-public assistance family with earnings of $7,000 eligible at this income level only for a small amount of medical assistance, would be left with a disposal income plus benefits of only about $5,800. Or put another way, the ordinary family would need gross earnings of about $10,000 to achieve the same standard of living as the female-headed AFDC family earning $7,000. If the female-headed family has earnings of $7,000 and is also benefiting from housing subsidies and has one child in a day care program, its disposable income plus benefits amounts to $11,300 which is equal to gross earnings of almost $15,000.

Bernstein, et al., supra at 146-47.

The system thus exerts pressure on the families of the working poor—whether on or off Home Relief—to dissolve. It further acts as a deterrent to the formation of family units. A recent federally funded study of some 500 AFDC mothers in New York City confirms this. B. Bernstein and W. Meezan, The Impact of Welfare on Family Stability (Center for New York City Affairs, New School for Social Research: June 1975). Fourteen percent of those interviewed expressly stated that the decision to break up their family or relationship had been influenced by the availability of welfare. The authors estimate that even more respondents, roughly 21 percent, were actually influenced. Id. at 98-99. Most significantly, welfare benefits were found to influence a higher percentage of respondents and to play a more influential role relative to other factors causing break-up where the welfare grant was more than or equal to the husband's or man's earnings (approximately 44 percent of the cases). Id. at 79, 97. Previous or current receipt of welfare was also found to be clearly related to the extent to which the decision to separate is influenced by the availability of welfare. Id. at 80-81. Thus the working poor presently on Home Relief are subject to the greatest pressures to dissolve intact families. The $80 ceiling—by increasing disparities between Home Relief and AFDC—necessarily increases these pressures. As Bernstein and Meezan put it, "Such discrimination makes no sense in a society which places high value on the family and recognizes its essential contribution to economic and social well-being." Id. at 105.

F. Cost Savings

Reductions in program costs are the only conceivable benefit which may result from the work expense ceiling. As pointed out above, cost savings alone do not justify an otherwise arbitrary classification. But beyond this, the amount actually saved by this limitation on work expense deductions is entirely problematic. The state has informed the court that the Department of Social Services has no figures from which to estimate the financial impact of eliminating the ceiling. Moreover, a major recent economic analysis of the cost of welfare programs strongly suggests that the ceiling is increasing the expense of the Home Relief program by discouraging recipients from earning additional income. See S. A. Rea, Jr., "Trade-Offs Between Alternative Income Maintenance Programs" in Studies in Public Welfare, Paper No. 13 (U.S. Government Printing Office, Washington, D.C.: February 18, 1974). Using a 1967 cross-sectional labor force survey linking wage rates and unearned income to hours worked, Rea found that, all other factors being equal, when projected work responses are taken into account, an income transfer program with a benefit loss rate of 67% produces lower budget costs than programs with benefit-loss rates of 100%, 50%, and 33%. Reducing tax rates below 67% swells costs by expanding eligibility into more densely populated higher income brackets. Raising the benefit-loss rate to 100% greatly increases in program costs because the extreme work disincentive causes a high percentage of recipients to reduce earnings. Id. at 50.

To some extent earnings reductions can be policed by rigorous enforcement of section 131(5) of the Social Services Law with its registration and employment requirements. But that statute cannot reach such rational behavior as choosing a lower paying job, avoiding optional overtime, refusing a second job, or preventing other family members from taking paid employment. See Bernstein et al., supra at 146.

Certainly, some positive economic incentives may flow from the ceiling. Some families may be able to reduce transportation expenses by moving closer to work; others may be influenced to use cheaper day care facilities. But these are the exceptions; all the other expenses enumerated in 18 NYCRR § 352.19 are outside the employee's control, except those directly related to income. For the average worker reducing earnings is the only available means of reducing work expenses.

If New York had been interested in eliminating unnecessary work expenses, it could have applied individual ceilings to the few items subject to employee control (as is presently the case with meals). That the statute applies to all expense items, including involuntary payroll deductions, unmistakably stamps it as a wholly arbitrary means of reducing benefits.

G. Summary

The inability of poor people to earn incomes above the subsistence level is, in part, a function of inadequate education and training, discrimination, and structural unemployment within the economy. Poor families, like middle-class and rich families, break up because of emotional stress, drug addiction, alcoholism and other difficulties. Pressures associated with poverty add to burdens on the less affluent. But, as the studies demonstrate, the welfare system in general and the work expense ceiling in particular contribute significantly to inducing people not to work and to destroying families.

The working poor probably represent that segment of our society most victimized by present welfare policies. Data collected recently by the University of Michigan Survey Research Center demonstrate that among poor families with children, those headed by able-bodied, working-age males are least likely to receive transfer payments under any program and most likely to remain poor if they do receive benefits. M. C. Barth, G. J. Carcagno, J. L. Palmer, Toward an Effective Income Support System: Problems, Prospects, and Choices 24-29 (1974). Employable recipients are most subject to pressures to misreport income and to accept a cash-paying job—possibly in an illicit occupation—rather than one covered by Social Security. See Introduction, Income Transfer Programs: How They Tax the Poor in Studies in Public Welfare, Paper No. 4, prepared for Subcommittee on Fiscal Policy, Joint Economic Committee vi-vii (Washington, D.C.: Government Printing Office, April 18, 1974). Employed recipients routinely suffer work disruptions and lost pay in order to cope with welfare bureaucracy. See New York Times, July 14, 1976, at 41 col. 1.

In strict terms, these observations do not bear on the narrow constitutional issue before us. It is enough that the challenged statute arbitrarily discriminates between applicants and recipients for no purpose other than possible short term reductions in program cost. But the studies do underscore the urgency of the question. We are dealing with real human deprivation in its most basic dimensions.

Discrimination against the working poor and against intact families transcends the narrow question of the validity of the work expense ceiling. It is at the heart of all serious discussions of welfare reform. See generally M. C. Barth, G. J. Carcagno, J. L. Palmer, Toward an Effective Income Support System: Problems, Prospects and Choices (1974); H. Aaron, Why is Welfare So Hard to Reform? (1973); D. P. Moynihan, Politics of a Guaranteed Income (1973); V. and V. Burke, Nixon's Good Deed, Welfare Reform (1974); New York Temporary State Commission to Revise the Social Services Law, Recommendations made to the 94th Congress and the 1976 New York State Legislature on Public Welfare Programs. The authors of the Bernstein-Shkuda-Burns study stress that inequities run throughout all the interrelated welfare programs. They conclude with the following indictment of the system:

In summary, the present package of social programs does not provide equity among the different types of families on public assistance or between those who are and those who are not on public assistance; nor does it produce incentives to increase income for those whose earnings potential may be limited.

It must also be stressed that the bewildering variety of criteria for eligibility, the differences in the definition of income and of income disregards for various programs, and the varying procedures for verifying income, resources, and other aspects of eligibility—some strict and some loose—fail to assure either that everyone will understand what he is entitled to receive or that those not entitled to benefits will in fact be denied them.

Finally, it must be noted that the variety of policymaking bodies involved at the Federal, State and local levels, each with some authority to determine the rules which will govern the criteria for eligibility and the nature of the administration, has led to serious inconsistencies in the rationale for deciding who should benefit and to what degree from various programs.

Bernstein et al., supra at 147.

V. Conclusion

The present welfare system is shot through with inequities and irrational distinctions. Cf., e. g., Liebman, The Definition of Disability in Social Security Income: Drawing the Bounds of Social Welfare Estates, 89 Harv.L.Rev. 833 (1976). Yet, although the system badly needs reform, changing the status quo, with all its entrenched beneficiaries, will not be easy. Cf. Tribe, Structural Due Process, 10 Harv.Civ. Lib.L.Rev. 269, 315 (1975).

Welfare issues involve federal, state and municipal laws and bureaucracies, enormous fiscal implications, conflicting pressure groups, and a lack of consensus as to basic purposes. Cf. I. Garfinkel, An Overview Paper in M. C. Barth, G. J. Carcagno, J. L. Palmer, Towards an Effective Income Support System 153-160 (1974). Reforming the system is thus primarily a task for Congress and the state and municipal legislatures, not the courts.

While continuing legislative failure to eliminate deep-rooted inequities may constitute a violation of constitutional rights, we should avoid, if at all possible, placing the process under judicial sanction. Cf. Robinson v. Cahill, 67 N.J. 333, 339 A.2d 193 (1975). But the courts should not remain silent where, as here, no rational justification consistent with any conceivable state policy can explain a blatant discrimination. An injunction against future application of the work expense ceiling is an entirely appropriate remedy for the violation of plaintiffs' rights to equal protection under the laws, fully consistent with the limited judicial role prescribed in Dandridge. This relief would terminate a particularly harsh inequity while leaving the complex task of restructuring the welfare laws to the appropriate legislatures.

I would grant plaintiffs relief.

NOTES

[1] Plaintiffs' Supplementary Memorandum of Law In Support Of Motions For Preliminary Injunction, Formation of Three-Judge Court and Class Certification And In Opposition to Defendants' Berger And New York State Dep't of Social Services Motion To Dismiss, p. 25.