475 F.2d 449

82 L.R.R.M. (BNA) 2986

GRINNELL CORPORATION, Plaintiff-Appellant,
v.
Mary C. HACKETT, Director of the Department of Employment
Security of the State of Rhode Island and John J. Affleck,
Director of the Department of Social and Rehabilitation
Services of the State of Rhode Island, Defendants-Appellees,
The Chamber of Commerce of the United States of America and
The Greater Providence Chamber of Commerce, Intervenors, Appellants.

Nos. 72-1275, 72-1276.

United States Court of Appeals,
First Circuit.

Heard Nov. 9, 1972.
Decided March 15, 1973.

George M. Vetter, Jr., Providence, R. I., with whom William R. Powers III, Providence, R. I., David D. McKenney, West Barrington, R. I., and Hinckley, Allen, Salisbury & Parsons, Providence, R. I., were on brief, for Grinnell Corporation.

Lawrence M. Cohen, Chicago, Ill., with whom Milton Smith, Denver, Colo., O. F. Wenzler, Washington, D. C., Jeffrey S. Goldman, Chicago, Ill., Guy J. Wells, Providence, R. I., Gerard C. Smetana, Chicago, Ill., Lederer, Fox & Grove, Chicago, Ill., and Gunning, LaFazia, Gnys & Seyla, Inc., Providence, R. I., were on brief, for The Chamber of Commerce of The United States of America and The Greater Providence Chamber of Commerce.

Warren H. Pyle, Boston, Mass., with whom Angoff, Goldman, Manning, Pyle & Wanger, Boston, Mass., were on brief, for intervenor, United Steelworkers of America, AFL-CIO.

W. Slater Allen, Jr., Asst. Atty. Gen., with whom Richard J. Israel, Atty. Gen., was on brief, for defendants-appellees.

Before COFFIN, Chief Judge, ALDRICH and McENTEE, Circuit Judges.

COFFIN, Chief Judge.

1

This case brings before us again the complex question whether state financial aid to striking workers, here unemployment compensation benefits, is preempted by the federal labor statutes and policy. We initially wrestled with that preemption question in a case involving state and federal welfare benefits made available to strikers. ITT v. Minter, 435 F.2d 989 (1st Cir. 1970), cert. denied, 402 U.S. 933, 91 S.Ct. 1526, 28 L.Ed.2d 868, reh. den., 404 U.S. 874, 92 S.Ct. 27, 30 L.Ed.2d 120 (1971). See Note, Welfare for Strikers: ITT v. Minter, 39 U.Chi.L.Rev. 79 (1971).1

2

On March 20, 1972, after unsuccessful negotiations for a new collective bargaining agreement between the plaintiff and Local 4756 of the United Steelworkers of America, all 585 of the plaintiff's employees went on strike. Some of them applied for and received welfare assistance, both state General Assistance, Sec. 40-6-8 General Laws of Rhode Island (R.I.G.L.) (1971 Pocket Supp.), and federally assisted AFDC-U benefits (aid to families with unemployed fathers), R.I.G.L. Sec. 40-6-7, 42 U.S.C. Sec. 607; and food stamps, 7 U.S.C. Secs. 2011-25.2 As of the week ending May 13, 1972, 404 of the plaintiff's employees received unemployment compensation benefits, R.I.G.L. Sec. 28-44-1 et seq., under a special provision which permits payments of such benefits, after the general one-week waiting period (Sec. 28-44-14) and a special additional six-week period, to individuals "unemployed because of a strike", R.I.G.L. Sec. 28-44-16.

3

On May 15, plaintiff filed the complaint in this case, seeking a preliminary and permanent injunction against the defendants Hackett and Affleck, who are the directors, respectively, of the Departments of Employment Security and Social and Rehabilitation Services of the State of Rhode Island. The defendants moved to dismiss for failure to state a claim upon which relief could be granted and subsequently filed an answer alleging failure to state a claim, lack of conflict between the federal and state statutes, and failure to join an indispensable party, the local union. The United Steelworkers of America, AFL-CIO, and the Chambers of Commerce of the United States and Greater Providence sought and were granted leave to intervene.

4

On May 30 and 31, the district court held a hearing on the motion for preliminary injunction at which the plaintiff presented three witnesses, two of whom, Armand Thieblot and Ronald Cowin, were co-authors of a study entitled "Welfare and Strikes: The Use of Public Funds to Support Strikers". The book, which was also introduced as an exhibit, was explicitly an attempt to answer some of the empirical questions which we indicated in Minter a court would have to consider in deciding the preemption issue. An affidavit by Herbert Northrup, the supervisor of the project at the University of Pennsylvania School of Finance which led to the book, was also introduced as an exhibit. Defendants presented two brief witnesses, one of whom was the defendant Hackett, and one exhibit, "Handbook of Labor Statistics 1971" published by the United States Department of Labor.

5

In its opinion, issued on June 15, the district court attempted to discern whether under Minter it could consider evidence as to the impact of unemployment benefits on federal labor policy and, if so, whether such evidence could overcome the state interest supporting the unemployment compensation scheme. It initially quoted at length the relevant portions of our opinion in Minter, which had come to us on a denial of a preliminary injunction on the pleadings. We there first noted that in labor preemption cases involving an alleged tangential frustration of national labor policy, rather than a direct overlap of state and federal regulations of labor activities as in San Diego Building Trades Council v. Garmon, 359 U.S. 236, 79 S.Ct. 773, 3 L.Ed.2d 775 (1959), "a balancing process seems called for . . . in which both the degree of conflict and the relative importance of the federal and state interests are assessed." 435 F.2d at 992. After noting our hesitance at making any a priori judgments regarding either the extent of conflict or the relative strengths of the federal and state interests, we described some of the questions of general scope which a court would have to consider, as to both issues, if a case went beyond the pleadings. We then stated that

6

"This very catalogue of data relevant to a macrocosmic weighing, which a court, if called upon would have to undertake, indicates the preferable forum to be the Congress." 435 F.2d at 993-994,

7

and noted the strong likelihood that Congress was aware of this problem. We concluded:

8

"In sum, wholly apart from the inadquacy of the evidence before the district court, we have substantial doubt that a significant frustration of federal collective bargaining policy is effected by the granting of welfare benefits to indigent strikers or that, even so, the state interest is so insubstantial compared to the federal interest that Congress must be supposed to have deprived the state of such power to serve that interest." 435 F.2d at 994.

9

The district court here, after quoting these and other passages, found "no distinctive difference" between the Rhode Island unemployment scheme and the Massachusetts welfare scheme upheld in Minter; ruled that the former's purpose was "to protect citizens against economic vicissitudes" and was "hardly abandoned" by the six-week waiting period; noted that "[i]t also appears that many of the Grinnell strikers would be receiving welfare were they not receiving unemployment compensation", that Congress was not unaware of the problem, and that the state interest in "the well-being of its unemployed and ultimately for the health of the local community" was substantial. The court then held that the preemption issue was not justiciable but that in any case "The state's interest is so substantial that this Court will not conclude that Congress has excluded such state action." Grinnell Corp. v. Hackett, 344 F.Supp. 749, 753-754 (D.R.I.1972). For these reasons, it denied the preliminary injunction, granted the motion to dismiss and made no findings on the issue of infringement of the federal collective bargaining process. The plaintiff and the two Chambers of Commerce appealed from the subsequent order to that effect.3

10

We are unable to affirm the district court on either ground of decision. First, we continue to believe, as we indicated in Minter, that the preemption issue is justiciable, albeit complex. Moreover, for reasons that we will develop more fully later, on the present record we are not able to determine whether the state interest in unemployment compensation is the same, or as substantial, as the state interest in welfare, as the district court seemed to believe, and we have substantial doubts whether that issue need even be reached. At the same time, we have serious doubts whether the existing record is sufficient to meet the plaintiff's burden of showing the probability of establishing that the Rhode Island unemployment compensation program substantially disturbs the collective bargaining process. Since, however, it is not for us to make findings of fact, we vacate the denial of a preliminary injunction. We also reverse the dismissal of the complaint for failure to state a cause of action, being unable to rule categorically that no evidence could be presented which would support the claim, see Dow Chemical Co. v. Taylor, supra, n. 3 (refusing to dismiss complaint alleging federal preemption of unemployment benefits to strikers). Since a remand is necessary, and further evidence may be presented, we will discuss the issues in the order that we believe a court should generally follow when faced with this question. After considering justiciability, we will examine whether "Congress has unmistakably ordained" that states not exercise power in this area. Florida Lime & Avocado Growers, Inc. v. Paul, 373 U.S. 132, 142, 83 S.Ct. 1210, 10 L.Ed.2d 248 (1963); Armour & Co. v. Ball, 468 F.2d 76 (6th Cir. 1972). Finding no such clear manifestation, we will then comment on the nature of the infringement issue, the inadequacy of the present record and suggest possible other evidence that might be more relevant and persuasive. Finally we will consider the state interests and the factual inquiries that might be material in that regard.

JUSTICIABILITY

11

Justiciability is, of course, as Baker v. Carr, 369 U.S. 186, 82 S.Ct. 691, 7 L.Ed.2d 663 (1962), taught, distinguishable from subject matter jurisdiction, which was here properly not disputed, a civil action wherein the matter in controversy exceeds $10,000 arising under the Constitution and laws of the United States, 28 U.S.C. Sec. 1331, in particular an Act of Congress regulating commerce, 28 U.S.C. Sec. 1337. Justiciability concerns the propriety of judicial decision in light of the posture of the controversy or, as here, the distribution of functions among the coordinate branches of the federal government. Two formulations of nonjusticiability arguably applicable here are "a lack of judicially discoverable and manageable standards for resolving [the case]; or the impossibility of deciding without an initial policy determination of a kind clearly for nonjudicial discretion", 369 U.S. at 217, 82 S.Ct. at 710. The district court interpreted Minter as saying that "the issue is one for decision by the Congress, not the Courts." 344 F.Supp. at 754. Although we did indicate there and still believe that the complexity and generality of the issue made Congress "the preferable forum", 435 F.2d at 994 [emphasis added], we did not say that the questions were therefore beyond the ken of the courts. Indeed, we spoke frequently of what "a court would" or "must" do "if called upon" by appropriate evidence. Id. at 993. Moreover, our sketching of the appropriate analysis, with relevant citations to other preemption cases, suggested that judicially manageable standards have been discovered by the Supreme Court and that courts are competent to determine, solely by the exercise of judicial judgment upon relevant evidence, whether Congress' initial policy determinations or the nature and effect of the federal and state policies at issue require a finding of preemption. Finally the very fact of our decision on the merits, and the Supreme Court's many decisions on the merits in labor preemption cases, even of the non-Garmon variety, see e.g., Teamsters Local 20 v. Morton, 377 U.S. 252, 84 S.Ct. 1253, 12 L.Ed.2d 280 (1964), indicate that the issues are justiciable. Although it is possible that a complete evidentiary record, along the lines suggested herein, might ultimately convince us that judicial fact-finding and decision-making are unmanageable or impossible in unusual preemption cases, such as this, see discussion infra, we cannot, on the present record and the present state of our knowledge, say that as to the issues here, unlike all other preemption issues, Congress is not just the preferable, but the only forum.

12

Before leaving justiciability, we feel compelled, in light of the position taken by appellee's counsel at argument, to reaffirm our holding in Minter that a case such as this is not moot, although the particular strike which prompted the suit and the payment of benefits challenged have, as here, terminated by the time of appeal. We are unable to say that the issue will not recur with regard to this very plaintiff-one strike of the length requisite to raise the issue has now taken place at its plant and nothing in the record suggests the unlikelihood of future repetition. See General Dynamics Corp. v. Local 5, Industrial Union of Marine and Shipbuilding Workers of America, AFL-CIO, 469 F.2d 848 (1st Cir. 1972). Moreover, the intervening Chambers of Commerce, who have filed their own complaint for declaratory and injunctive relief, represent many other employers in Rhode Island, as to at least some of whom the issue is certain to recur.

CONGRESSIONAL INTENT

13

If Congress has made clear its intent, either to preempt or not to preempt, no further inquiry is required. Yet, if the intent is not unmistakable, and evidence and legal analysis must be undertaken, Congressional awareness of the problem and its ability to resolve it finally and specifically, may, as we said in Minter, in a closely balanced situation, be sufficient justification for upholding the state action. We find that although the existing legislative record is not sufficiently clear to establish Congressional intent either way, it strongly indicates Congressional awareness, the availability of opportunities to act, and Congressional action in closely related matters which would prove relevant should the evidence on infringement and state interest be closely balanced.

14

In July 1933, the Federal Emergency Relief Administration (FERA) ruled that it would not "attempt to judge the merits of labor disputes" but would treat unemployed strikers like all other unemployed persons for purposes of relief. Brown, Public Relief 1929-39, 270 (1940). This policy became a bone of contention during the textile strike of September 1934. Bernstein, The Turbulent Years: A History of the American Worker 1933-41, 307 (1971). Congress could hardly have been unaware of this policy when in 1935 it passed in close succession both the National Labor Relations Act, 49 Stat. 449, and the Social Security Act, 49 Stat. 620, Title IX of which, 49 Stat. 639-645, is now the Federal Unemployment Tax Act, 26 U.S.C. Secs. 3301-09. Indeed, in detailing the requirements for state laws that would exempt employers from the federal unemployment tax, Congress specifically established three conditions to insure the compatibility of state unemployment compensation laws with the then brandnew labor statute-unemployment compensation could not be denied to an otherwise eligible individual for refusing to accept new work (1) if the position offered was vacant due directly to a labor dispute, (2) if the wages for the offered job were substantially less favorable than those prevailing locally for comparable work or (3) if a condition of the offered employment was membership in a company union or resignation from a bona fide union. Sec. 903(a)(5), 49 Stat. 640, 26 U.S.C. Sec. 3304(a)(5).4 Certainly, had it thought it necessary to preserve its labor policy, Congress could also have required state laws to bar payment of benefits to strikers. Moreover, pursuant to Sec. 3304(a) the Social Security Board, presumably aware of the Congressional intent, approved, within the next few years, four state laws which allowed payments to strikers. Ch. 468, Sec. 504(2)(b), 1-1935 N.Y.Laws 1033, as amended N.Y.Labor Law Sec. 592(1) (McKinney's Consol.Laws, c. 31, 1965); 1935-36 R.I.Pub.Laws 848 (1936), as amended R.I.G.L. Sec. 28-44-16; Act No. 1, Sec. 401(e), 2-1937 Pa.Laws 2914 (2d Extra Sess.1936) (repealed 1947), see 43 Pa.Stat.Ann. Sec. 802(d); No. 164, Sec. 4(d), 1938 La.Acts 394 (repealed 1946), see 23 La.Rev.Stat.Ann. Sec. 1601(4).

15

In 1947, the Hartley bill, passed by the House, provided that a striker who accepted unemployment compensation benefits would no longer be considered an "employee" under the National Labor Relations Act, Sec. 2(3), H.R. 3020, 80th Cong., 1st Sess. (1947), and thus would lose all rights under the Act, because, as the committee report indicated, such benefits were "a perversion of the purposes of the social security laws". H.R.Rep.No.353, 80th Cong., 1st Sess. 12 (1947). In conference, however, the provision was dropped without explanation. Conf.Rep.No.510, 80th Cong., 1st Sess. (1947), 1947 U.S.Cong.Service 1135, 1137-39. Given this silence, the enormous political controversy surrounding the Taft-Hartley Act, and the consequent need for compromises, perhaps unreasoned or hasty, we cannot read this deletion and the subsequent approval of the conference bill as specific congressional resolution of the problem.

16

More recently, in 1969, President Nixon, in a message to Congress, specificially proposed a requirement that states deny unemployment benefits to strikers. Message from President Nixon, Hearings Before the House Committee on Ways and Means on H.R.12625, 91st Cong., 1st Sess., at 12 (1969). The Committee, however, deleted the provision in creating a substitute bill (H.R. 14705), H.R.Rep. No. 612, 91st Cong., 1st Sess. (1969), because, as Chairman Mills explained briefly on the floor of the House in the course of discussing other aspects of the bill:

17

"We have tried to keep from prohibiting the States from doing the things the States believe are in the best interest of their people. . . . For example, there are two States . . . which will pay unemployment benefits when employees are on strike. . . . [I]f the State wants to do it we believe they ought to be given latitude to enable them to write the program they want." 115 Cong.Rec. 34106 (1969).

18

The House passed the substitute bill without amendment or further discussion of this deletion, 337 to 8, 115 Cong.Rec. 34110-11(1969). The following year, the Senate considered the House bill which made, of course, no mention of the striker benefits issue. In the course of explaining a provision preventing states from cancelling an individual's benefit credits because of an act which disqualified him under state law from receipt of benefits for the subsequent period of unemployment, the Senate Finance Committee Report stated:

19

"It would not prevent a State from specifying the conditions for disqualification such as, for refusing suitable work, for voluntary quitting, for unemployment due to a labor dispute in the worker's plant, etc. . . . It would not preclude disqualifications which only postpone the receipt of benefits for a specified or flexible number of weeks." S.Rep.No.752, 91st Cong.2d Sess. (1970); 1970 U.S. Code Cong. & Admin.News, pp. 3606, 3627.

20

Although during the course of the Senate debate, 116 Cong.Rec. 10573-90 (1970), this provision preventing cancellation of credits was described, id. at 10575, no mention was made of either the state disqualifications for labor disputes or of the striker benefits issue. Neither the Conference Report, Conf. Rep.No.1037, 91st Cong. 2d Sess. (1970), 1970 U.S.Code Cong. & Admin.News, p. 3646, nor the subsequent floor debates-116 Cong.Rec. 25608-17 (House), 27305-23 (Senate) (1970)-made any mention of the issue. We are aware of the importance and power of the House Ways and Means Committee and of the significance of its Chairman's statements on bills within its jurisdiction. Yet, in the absence of any floor amendment or debate related to that particular change in either house, and of any real consideration of the issue at any level in the Senate, we cannot take the House Committee's deletion, and Congress' subsequent approval of the substitute bill, as a clear indication of the intent of the entire Congress not to preempt unemployment payments to strikers.

21

Congress has, on the other hand, shown its ability to act specifically and directly in other, closely related areas. In the Railroad Unemployment Insurance Act, 52 Stat. 1094 (1938), as amended 45 U.S.C. Secs. 351-367, Congress explicitly provided that only those engaged in unlawful strikes were disqualified from unemployment benefits. 45 U.S.C. Sec. 354(a-2)(iii). A recent Administration bill to repeal this provision, S. 560, 92nd Cong., 1st Sess. (1971), did not pass. Similarly, in 1971, Congress amended its original scheme in which the states were allowed to set eligibility standards for the Food Stamp Act, Sec. 5(b), 78 Stat. 703, 704 (1964), and required the Secretary of Agriculture to set "uniform national standards of eligibility." 7 U.S.C. Sec. 2014(b) (1972 Supp.) Although it specifically required a national standard disqualifying households with able-bodied adults who refuse to register or accept employment, Congress added a proviso:

22

"Refusal to work at a plant or site subject to a strike or a lockout for the duration of such strike or lockout shall not be deemed to be a refusal to accept employment." 7 U.S.C. Sec. 2014(c) (1972 Supp.).

23

The stated reason for this provision and rejection of a provision disqualifying strikers was Congress' determination not "to take sides in labor disputes", H.R.Rep.No. 1402, 91st Cong., 2d Sess. (1970), 1970 U.S.Code Cong. & Admin. News, pp. 6025, 6035, a phrase reminiscent of the FERA policy. Likewise, in the AFDC-U program, Congress, after initially leaving the determination of standards to the state, Sec. 407, 75 Stat. 75, insisted in 1967 that the Secretary of Health, Education and Welfare determine the standard of unemployment, 42 U.S.C. Sec. 607(a), (b)(1)(A). In Davidson v. Francis, 409 U.S. 904, 93 S.Ct. 223, 34 L.Ed.2d 168 (1972), the Supreme Court affirmed the judgment of a three-judge court, 340 F.Supp. 351 (D.Md.1972), which construed literally the language of the subsequent regulation, 45 C.F.R. Sec. 233.100, holding that "any father" included one working less than the requisite hours because of a strike and finding no preemption problem with such aid, citing Minter. Thus, presently federal law also requires payment of AFDC-U benefits to strikers in the 24 states participating in that program.5

24

The existence and interpretation of these provisions in analogous programs might be read as indicating a broad Congressional intent not to preempt payments to strikers under public assistance statutes. Yet, given the ability of Congress to articulate that intent in other programs, one might also infer that its silence in the unemployment compensation statute was indicative of a contrary intent. That, too, however, is not a reasonable inference, given the pre-1935 policy, the 1935 Acts and the rejection of explicitly prohibitory legislation in both 1947 and 1969. The most that can fairly be said, in the face of this legislative record, is that Congress has been and presently is aware of the problem, has had the opportunity to resolve it, and has acted in closely analogous circumstances.

25

Since unambiguous Congressional intent in lacking, a court must consider the following questions: does the payment of unemployment compensation benefits to strikers "palpably infringe" upon federal labor policy? Southern Pacific Co. v. Arizona ex rel. Sullivan, 325 U.S. 761, 766, 65 S.Ct. 1515, 89 L.Ed. 1915 (1945), and is the state interest in cushioning the impact of unemployment stronger than the federal interest in untrammelled collective bargaining? Although a negative answer to the first question or an affirmative answer to the second would, strictly, dispose of the case, a district court would be well advised to consider both issues, as it considers both patent validity and infringement in a patent case. Naturally, if palpable infringement is found, the strength of the state interest must be explored; and if the state interest is first approached and found not more compelling than the federal interest in free collective bargaining, then the issue of palpable infringement must be inquired into. Should a court, however, find infringement but also that the state and federal interests are closely balanced it could properly decline to find preemption because of the historical and continuing Congressional awareness, opportunity, and demonstrated willingness to act in this field.

INTERFERENCE WITH LABOR POLICY

26

Since the district court has explicitly declined to make any findings on the issue of infringement, and in any event on remand this issue remains open, we think it well to alert court and counsel to critical problems of relevance posed by the study and testimony thus far put into evidence.

27

We note initially the need to define precisely the infringement question relevant to unemployment benefits. The Rhode Island unemployment compensation for strikers statute under attack here, like its sole companion in the state of New York,6 does not provide strikers with benefits from the start. Rather it imposes, in addition to the general waiting period for all unemployed claimants, Sec. 28-44-14, an additional six-week waiting period for striking claimants, Sec. 28-44-16.7 It is thus fundamentally different from state welfare programs and the federal food stamp program which provide for payments immediately upon a showing of need.8 The relevant question then in determining this statute's impact upon the federal collective bargaining policy is whether the receipt or expectation of receipt of unemployment compensation benefits after seven weeks of a strike causes workers to stiffen bargaining demands beyond those they would have made without such benefits, to strike in the first instance when they would otherwise have settled, or to continue a strike for a longer period than they otherwise would.9

28

To date, plaintiff has not addressed this question directly. Rather it has relied on an extensive empirical study of the payment of welfare benefits and food stamps, available from the moment of need, to strikers in thirteen strikes. In only four were any unemployment benefits paid, in each case in New York state. Whatever support this massive body of data would provide for a finding regarding the impact of receipt of welfare and food stamps on strike length, and as we note below that would seem minimal at best, it cannot support any finding regarding the impact of unemployment benefits not received until after seven weeks, in the absence of some effort to differentiate the separate impacts of the different forms of aid.

29

Only a small proportion (17 per cent) of strikers in all strikes studied received welfare aid roughly equivalent to unemployment benefits.10 Our computations indicate that an even smaller proportion (7 and 13 per cent in the only two strikes for which relevant figures were provided) received unemployment benefits. It would not seem to us tenable to assume, a priori, that such a small percentage of the total strike force would have a determinative impact on a strike-related decision of a majority of workers.

30

Moreover, as the study itself admits, the exceedingly long strikes used in the study, during which strikers in relevant states could qualify for unemployment benefits, are highly atypical. As tables in the defendant's Handbook establish, the national average duration of work stoppages between 1936 and 1969, when unemployment benefits were available to strikers in Rhode Island and New York, ranged from a low of 5 days (in 1943) to a high of 25.6 days (in 1947). In the period from 1959-69, the percentage of work stoppages nationwide lasting 60 days or more never exceeded 11.7 per cent.

31

Aside from the irrelevance of welfare figures and the small proportion of persons and strikes affected, the present record suffers from a fundamental defect. It provides no support for a causal relationship between the receipt of benefits, which unions obviously desire and often actively seek, and longer, costlier strikes. It may be, as the senior author of the study testified, that "there is no way of really determining whether welfare use makes a strike longer or costlier". Yet the record lacks even a crude form of what we assume would be the most relevant and probative type of evidence-statistical comparisons of the length and cost of strikes in states granting unemployment benefits (Rhode Island and New York) and the length and cost of strikes of similar size in similar industries in other states not granting such benefits. Such general statistical data would at least provide some hard primary evidence on which to start analysis and base broad opinion surveys or other particularistic empirical approaches. Their weight would depend on the extent to which they considered and controlled for other factors potentially responsible for any disparities-such as the size and nature of the industry, the length and strength of union representation, past labor-management relations and prior strike history, the relative wage and benefit structure under the expiring agreement, and local cost-of-living or other economic or structural peculiarities.

32

The only attempt to provide the missing causal link is the transcription, presumably partial, of six interviews with labor, management and welfare representatives in two particular strikes (out of a total of 60 interviews said to have been conducted). While we find these interviews oversimplistic and conclusory, and the answers accordingly wholly expectable, a more representative and sophisticated sampling of experience and attitudes directed at the effect of the actual or expected receipt of delayed unemployment benefits might be helpful to a court seeking to decide the relevant question of infringement.

STATE INTEREST

33

If a court were to find, based on relevant and persuasive evidence, that payment of unemployment benefits significantly frustrated the federal policy of collective bargaining, it would then have to consider whether the unemployment scheme represented a compelling state interest "so deeply rooted in local feeling and responsibility that . . . we could not infer that Congress had deprived the States of the power to act." Garmon, supra, 359 U.S. at 244, 79 S.Ct. at 779. Here the district court found the interest so substantial as to override any frustration of labor policy, however significant. We are not prepared to so hold on the present record, especially given our doubts that the issue need be reached.

34

First, as we noted in Minter, the state interests in unemployment compensation and welfare do not appear on their face identical. Welfare payments, premised on a finding of need, are directed at alleviating real hardship-serious threats to the physical well-being, indeed in some cases survival, of individuals. Unemployment benefits, in contrast, are conditioned on a certain minimal amount of prior employment and earnings in a specified period, R.I.G.L. Sec. 28-44-11, computed as a percentage of past earnings, within fixed outer limits, Sec. 28-44-6 (additional amounts of $5.00 per week are paid for each dependent child under 18, up to a maximum of four dependents), and limited in duration to a certain percentage of prior weeks worked, again with a fixed maximum (26 weeks), Sec. 28-44-9. The amount of assets on hand or the nature of individual or family need are irrelevant. The state interest would then appear to be, as the statute's declaration of policy, Sec. 28-42-2, indicates, the avoidance of "economic insecurity", or put another way, insurance against the temporary disruption of a flow of income.

35

Indeed, as to unemployment compensation for strikers, the state interest would appear somewhat narrower. The provision for such payments, Sec. 28-44-16, by imposing an extra six-week waiting period in addition to the generally applicable one-week period, obviously does not insure an uninterrupted flow of income. Rather it seems aimed at preserving the standard of living and meeting the obligations previously undertaken with the reasonable expectation that the prior flow of income would continue. It is common knowledge (perhaps even judicially noticeable, if that were necessary) that consumer payment obligations in our increasingly credit-dependent society are monthly. See Sniadach v. Family Finance Corp., 395 U.S. 337, 89 S.Ct. 1820, 23 L.Ed.2d 349 (1969); Fuentes v. Shevin, 407 U.S. 67, 92 S.Ct. 1983, 32 L.Ed.2d 556 (1972). One month's bills can usually be met from cash on hand (frequently derived from the last pay-check which is often distributed after the commencement of the strike) or delayed by discussions with understanding and rational creditors. A second flood will, however, often breach the dike. Indeed, although probably not essential to decision,11 evidence might establish that more than two months' interruption of income-and hence credit payments-would produce substantial defaults, with consequent foreclosures, repossessions (garnishment, of course, being a worthless remedy when no income exists) and even bankruptcies. The state certainly has an interest, the quantum of which we need not now decide, in avoiding such economic and social dislocations.12

36

Nevertheless, although purely statutory analysis suggests differences in the state's interest in welfare and in unemployment compensation, we are not certain that in practice the interests do in fact diverge. The district court stated, in its discussion under the heading "Conclusions of Law" that "It also appears that many of the Grinnell strikers would be receiving welfare were they not receiving unemployment compensation." 344 F.Supp. at 753. The record before the court indicated that no persons received both unemployment compensation and AFDC-U, see also 42 U.S.C. Sec. 607(b)(2)(C)(ii); 45 C.F.R. Sec. 233.100(c)(1)(v)(b), and that although 69 strikers were initially found eligible for public assistance, only 16 were still receiving such aid as of the date of the hearing because "the remainder were found ineligible because of increased income due to receipt of Unemployment Compensation benefits." Had there been a finding of fact that a substantial percentage of striking workers in Rhode Island would, if denied unemployment compensation, qualify for public assistance, this would constitute a ground for concluding that in fact the state interest in unemployment compensation is substantially the same as its interest in welfare, the quantum of which we again need not decide now.

37

Finally, whether or not the evidence permitted a conclusion that the state interest served by welfare and unemployment was at bottom the same, a court could consider the secondary economic and social effects of unemployment payments or their absence. In Minter, we suggested at least two possible effects of financial aid-minimization of violence in labor disputes, insofar as caused by the bitterness attendant upon the hardships normally suffered by strikers, and avoidance of economic stagnation in local communities, by which we primarily meant the effect on local businesses of the sudden drop in consumer demand resulting from the abrupt and total termination of income for a substantial part of the community's families. The book, one author, and the project supervisor all indicate that the study, although showing incidents of violence in strikes where state aid was made available, did not provide any basis for a conclusion either way. The author and supervisor both stated, however, as a matter of common sense, that strikes in which assistance was provided to strikers might produce higher settlements with their generally attendant inflationary effects. The latter also suggested that longer strikes with costlier outcomes could make some employers non-competitive, thus forcing them to close, with the resulting long-term unemployment and local economic stagnation. Again, the complex nature of the issues suggests that empirical evidence and expert economic testimony as to the nature and scope of both the long-term impact on industry pricing and competitiveness and the short-term impact on consumer demand and the relationship between the two could assist courts in evaluating the substantiality of the state interest.

38

We are not unaware of the unusual nature of our insistence on an appropriate and detailed evidentiary basis for deciding the preemption issue, and in particular the issue of infringement of federal policy. We believe, however, that this is the unusual preemption case. For unlike Garmon labor preemption cases or non-labor preemption cases like Florida Lime, one cannot ascertain from the face of the statutes whether they affect the same activity. Similarly, although our statute, like that in UAW v. Russell, 356 U.S. 634, 78 S.Ct. 932, 2 L.Ed.2d 1030 (1958), unambiguously announces the state interest promoted, the practical overlapping of parallel statutes and the complex and uncertain effects of the state action suggest that we cannot simply apply the relevant legal test to the statutes on their face. Our efforts to create an adequate evidentiary record should not, then, be taken in any sense as a departure from, or attempt to distinguish, the governing law of preemption;13 they are merely attempts to apply that law to a most unusual circumstance.

39

For the reasons already described, we vacate the district court's denial of the preliminary injunction, reverse its grant of the defendants' motion to dismiss for failure to state a claim, and remand the case for further proceedings consistent with this opinion.

1

See also Note, Congressional Preemption in Labor Relations, 6 Suffolk U.L.Rev. 263 (1972); Note, Welfare Assistance to Strikers in Need: The Protestant Ethic Revisited, 67 Nw.U.L.Rev. 245 (1972); Lascaris v. Wyman, 340 N.Y.S.2d, 292 N.E.2d 667 (1972)

2

No questions regarding the welfare payments or food stamps are before us

3

In addition to the briefs of the parties, the court made use, after providing the parties with notice and an opportunity to respond, see Code of Judicial Ethics Canon 3A(4), of the briefs of the student contestants in the Harvard moot court finals competition. Although the case used there, entitled Agar Chemical Corp. v. Taylor, was fictitious, it was patterned roughly after Dow Chemical Co. v. Taylor, 57 F.R.D. 105 (E.D.Mich.1972), and contained many of the issues involved in this case

4

The Rhode Island statute contains, of course, these necessary provisions, R.I.G.L. Sec. 28-44-20, as well as the provision for payment to strikers, Sec. 28-44-16

5

The Secretary of Health, Education and Welfare is presently in the process of revising the definition. Also, a bill introduced in the last Congress, as an amendment to Administration's Family Assistance Plan, would have denied strikers such assistance. H.R. 6004, 92nd Cong., 1st Sess. (1971)

6

N.Y. Labor Law Sec. 592(1) (McKinney 1965)

7

New York law imposes effectively an eight-week waiting period-a general "four effective days" period, N.Y. Labor Law Sec. 590(9) (McKinney 1965), and a special additional seven-week period for strikers. Id. Sec. 592(1). In Michigan, strikers can apparently claim unemployment compensation benefits after obtaining, and being laid off from, interim employment. See Dow Chemical, supra n. 3

8

General Assistance-R.I.G.L. Secs. 40-6-5, 8, 22, Sec. 40-5-1 (1971 Pocket Supp.); N.Y. Social Services Law Secs. 131, 158 (McKinney 1972-73 Pocket Part), Lascaris, supra n. 1; see also Mass.Gen. Laws Annot. ch. 117, Sec. 1 (1973 Pocket Part). Food Stamps-7 U.S.C. Sec. 2014(b); 7 C.F.R. Secs. 271.1(s), 271.3. The study introduced in this case notes the availability of General Assistance from the date of demonstrated need but adds that most states attempt, after the federally mandated 30-day waiting period for AFDC-U, 42 U.S.C. Sec. 607(b)(1)(A), to transfer individuals to that federally assisted program. Such a transfer is advantageous to the individual involved because, among other reasons, it automatically includes his family in the state's Medicaid program, 42 U.S.C. Sec. 1396a(a)(10)

9

We have phrased these questions in general terms, although we are only dealing directly with the Rhode Island statute, because, as we noted in Minter, preemption decisions are to be made via macrocosmic rather than particularistic inquiries. Thus evidence as to the New York experience, with the questions rephrased to reflect the longer waiting period, and the Michigan experience, if sufficiently widespread, would be both relevant and significant

In addition, we are fully aware that the impact of subjective expectations is less susceptible to proof. We include reference to expectations not to indicate that adequate proof regarding actual receipts would be insufficient to establish infringement-which would not be correct-but rather to suggest that proof of infringement need not be limited to cases of actual receipt.

10

Although 60% of the strikers received food stamps, that form of aid is not comparable in value and hence in impact to unemployment benefits, as the stipulations in this case and the tables on national averages in the book's Appendix indicate. Here the average striker received $44.00 per month in food stamp values while the average unemployment benefit was $77.20 per week. That food stamps are complementary to, rather than competitive with, unemployment benefits is also suggested by the fact that the two benefits may be received at the same time. In contrast, the statutory prohibition against joint receipt of AFDC-U and unemployment benefits, 42 U.S.C. Sec. 607(b)(2)(C)(ii), and the similar amounts provided-in this case AFDC-U benefits were $280 per month-strongly suggest their comparability in value

11

Courts generally determine state interests from unchallenged state assertions, e. g., one-year residency requirements decrease the amount of welfare payments, or irrefutable policy concerns, e. g., the state has an interest in preventing voting frauds, rather than from evidence. Although we are not requiring evidence with regard to the state interests, we are not prepared to say that in this unusual kind of preemption case, see discussion infra, evidence might not be properly introduced and considered with regard to the substantiality, if not the very existence, of the state interests

12

We do not mean to intimate that we have found specific evidence in the legislative record or elsewhere that Rhode Island legislators had precisely these consumer problems in mind when, in 1936, they imposed this seven-week waiting period. Nor do we mean to exclude the possibility that the legislators specifically intended to force management to capitulate after seven weeks by aiding strikers. Bad motivation, however, will not vitiate otherwise sound legislation. Palmer v. Thompson, 403 U.S. 217, 91 S.Ct. 1940, 29 L.Ed.2d 438 (1971). Moreover, since the state interest in such legislation may change over time, McGowan v. Maryland, 366 U.S. 420, 81 S.Ct. 1101, 6 L.Ed.2d 393 (1961), and since the object of this inquiry is the present effect on both federal labor policy and a state's interests, a court is obliged to consider possible present, as well as historical, rationales

13

We are, however, aware of and impressed with the suggestions of Professor Cox for expanding the currently existing rule for non-Garmon labor preemption cases, which he suggests was enunciated in Morton, supra, into a single rule for all labor preemption cases. Cox, Labor Law Preemption Revisited, 85 Harv.L.Rev. 1337 (1972). We are, of course, not in a position to adopt or reject such a test. In any case, although he does not discuss the kind of preemption case we are here faced with, we believe that his formulation is similar to our own statement of the existing rule for non-Garmon preemption cases and that it would, in our unusual case, probably still require the kind of evidence we have called for