dissenting:
I respectfully dissent. The majority opinion fails to apply the principles set out in the recent decision of the Supreme Court in Hazelwood School District v. United States, — U.S. —, 97 S.Ct. 2736, 53 L.Ed.2d 768 (1977), and our own decision in Equal Employment Opportunity Commission v. Local 14, 553 F.2d 251 (2d Cir. 1977). These decisions cast substantial doubt on the existence of illegal discrimination by these unions; there is no question that they require a remand to fix the hiring goal at a more reasonable figure.
I.
The majority opinion suggests that the finding of discrimination and the fixing of the hiring goal are insulated from review by our prior holding in Equal Employment Opportunity Commission v. Local 638, 532 F.2d 821, 830 (2d Cir. 1976). To the contrary, the doctrine of law of the case does not bar us from reconsidering and correcting our prior erroneous holding.
It is hornbook law that an appellate court must apply the law as it stands at the time of its decision. Vandenbark v. Owens-Illinois Glass Co., 311 U.S. 538, 61 S.Ct. 347, 85 L.Ed. 327 (1941); United States v. Schooner Peggy, 5 U.S. (1 Cranch) 103, 2 L.Ed. 49 (1801). When an appellate court has previously passed on some of the questions presented, but remanded for reconsideration of others, the general practice is to avoid re-examination of the issues determined by the first appeal. However, this is not a question of judicial power, but of judicial economy. Banco Nacional de Cuba v. Farr, 383 F.2d 166, 178 (2d Cir. 1967), cert. denied, 390 U.S. 956, 88 S.Ct. 1038, 20 L.Ed.2d 1151 (1968). As Professor Moore notes, this doctrine of law of the case is not a barrier to “self-correction of judicial error.” IB Moore’s Federal Practice 10.404[1], at 401 (2d ed. 1974). Thus, a court is always free to exercise its discretion to reconsider its previous rulings in a case before it.1 Messenger v. Anderson, 225 U.S. 436, 444, 32 S.Ct. 739, 56 L.Ed. 1152 (1912) (Holmes, J.); Perrone v. Pennsylvania R. Co., 143 F.2d 168, 169 (2d Cir. 1944) (Frank, J.); White v. Higgins, 116 F.2d 312, 317 (1st Cir. 1940) (Magruder, J.); Higgins v. California Prune & Apricot Grower, Inc., 3 F.2d 896 (2d Cir. 1924) (L. Hand, D. J.).
When a change in controlling law intervenes between two proceedings in the same court, the judgment should be modified to conform to current law insofar as it requires a future course of conduct. This is true whether the change is the result of statutory amendment, System Federation v. Wright, 364 U.S. 642, 81 S.Ct. 368, 5 L.Ed.2d 349 (1961); Pennsylvania v. Wheeling & Belmont Bridge Co., 59 U.S. (18 How.) 421, 431-32, 15 L.Ed. 435 (1855), or an intervening decision by the same court, Davis v. United States 417 U.S. 333, 94 S.Ct. 2298, 41 L.Ed.2d 109 (1974) (intervening decision by same Court of Appeals); Hampton v. Graff Vending Co., 516 F.2d 100, 103 (5th Cir. 1975) (same). Thus, where a party, as here, is under a continuing injunction barring conduct once considered unlawful, but now permitted, it is an abuse of discretion not to conform the injunction to prevailing law. System Federation v. Wright, supra, 364 U.S. at 646-*3850, 81 S.Ct. 368 (Harlan, J.).2 See United States v. Swift & Co., 286 U.S. 106, 114-15, 52 S.Ct. 460, 76 L.Ed. 999 (1932) (Cardozo, J.).
Even though the challenged decree commands actions thought beneficent, these principles must be applied. Thus, in Pasadena City Board of Education v. Spengler, 427 U.S. 424, 96 S.Ct. 2697, 49 L.Ed.2d 599 (1976), the Supreme Court was faced with a decree meant to prevent re-segregation of a school district already integrated by court order. After the decree was entered, the school district did not appeal. Subsequently, the Supreme Court’s decision in Swann v. Charlotte-Mecklenberg Board of Education, 402 U.S. 1, 91 S.Ct. 1267, 28 L.Ed.2d 554 (1971), made it clear that the district court had exceeded its authority in ordering such relief. The defendants sought a modification of the injunction under Fed.R. Civ.P. 60(b)(5), which was denied, and the Court of Appeals affirmed this decision. The Supreme Court reversed. It held that law of the case did not insulate the decree from attack and that, insofar as it governed future conduct, the district court should have conformed it to current law. 427 U.S. at 437-38, 96 S.Ct. 2697.
There can be no doubt that Hazelwood and Local 14 significantly changed the law, delimiting for the first time the standard for selecting the labor force used as a benchmark in establishing a prima facie case of discrimination. Hazelwood, supra, — U.S. at —, 97 S.Ct. 2736; Local 14, supra, 553 F.2d at 254. The majority opinion as much as concedes their relevance. It states:
It is true that we have recently held that “where a significant number of union members come from outside the union’s geographic jurisdiction, the court must widen its sights; the appropriate reference area then should be that region from which the union draws its members.” EEOC v. Local 14, 553 F.2d 251, 254 (2d Cir. 1977). See also Hazelwood School District v. United States, — U.S. —, n.17, 97 S.Ct. 2736, 53 L.Ed.2d 768 (June 28, 1977). In Hazelwood and in Local 14, however, the court was concerned with a statistical basis for a finding of discrimination. Here we have no such problem, discrimination having been established by direct evidence of longstanding practices. See EEOC v. Local 638, 532 F.2d 821, 826.
553 F.2d at 2 (footnote omitted). This circular argument proves nothing at all. The discrimination that was “established by direct evidence” was premised upon the challenged test, which cannot possibly be its own justification. Hazelwood and Local 14 set out the test to be used in establishing discrimination, and conclusively demonstrate that the finding of liability in the earlier proceedings was based on incorrect standards. I am at a loss to see how a determination of liability under an errone*39ous test insulates itself from judicial scrutiny once the correct test is laid out by the Supreme Court. The majority should concede that we have no discretion to apply law of the case when the holding of the first appeal has been discredited by the Supreme Court. Zdanok v. Glidden Co., 327 F.2d 944, 951 (2d Cir.) (Friendly, J.), cert. denied, 377 U.S. 934, 84 S.Ct. 1338, 12 L.Ed.2d 298 (1964); Higgins v. California Prune & Apricot Grower, Inc., supra, 3 F.2d at 897 (2d Cir. 1924) (L. Hand, D. J.). Because neither the district court nor the first panel that heard this case had these decisions before it, the case should be remanded for reconsideration in the light of subsequent developments.
II.
In this ease, the plaintiffs established their prima facie case of discrimination by proving that the percentage of minority workers in the union was substantially less than the percentage of minority workers in the population. The district court, and now the panel majority, endorse this view. However, under our decision in Local 14, this finding resulted from the application of a clearly erroneous legal standard.
Title VII is addressed only to discrimination which has occurred since the effective date of the Act. Hazelwood, supra, — U.S. —, 97 S.Ct. 2736; International Brotherhood of Teamsters v. United States, 431 U.S. 324, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977). The district court, the original panel decision, and the majority here, however, looked to employment figures which included pre-Act membership in the union to establish a prima facie case of discrimination. In Hazelwood, the Court of Appeals concluded that a school district which had 1.8 percent minority teachers, in an area whose labor pool was 5.7 percent minority, had unlawfully discriminated. In vacating the decision, the Supreme Court stated:
The Court of Appeals totally disregarded the possibility that this prima facie statistical proof in the record might at the trial court level be rebutted by statistics dealing with Hazelwood’s hiring after it became subject to Title VII. Racial discrimination by public employers was not made illegal under Title VII until March 24, 1972. A public employer who from that date forward made all its employment decisions in a wholly nondiscriminatory way would not violate Title VII even if it had formerly maintained an all-white work force by purposefully excluding Negroes. For this reason, the Court cautioned in the Teamsters opinion that once a prima facie case has been established by statistical work force disparities, the employer must be given an opportunity to show “that the claimed discriminatory pattern is a product of pre-Act hiring rather than unlawful post-Act discrimination.” [— U.S. at —, 97 S.Ct. at 2742].
The record in this case showed that for the 1972-1973 school year, Hazelwood hired 282 new teachers, 10 of whom (3.5%) were Negroes; for the following school year it hired 123 new teachers, five of whom (4.1%) were Negroes. Over the two-year period, Negroes constituted a total of 15 of the 405 new teachers hired (3.7%). Although the Court of Appeals briefly mentioned these data in reciting the facts, it wholly ignored them in discussing whether the Government had shown a pattern or practice of discrimination. And it gave no consideration at all to the possibility that post-Act data as to the number of Negroes hired compared to the total number of Negro applicants might tell a totally different story.
— U.S. at —, 97 S.Ct. at 2743 (footnotes omitted). We reached the identical conclusion in Local 14, supra, 553 F.2d at 254.
It is by no means a futile gesture to allow this union a chance to produce such evidence. In Local 14 we overturned a finding of liability against another construction union whose jurisdiction was New York City. That union had a minority membership of 6.5 percent; however, when its post-Act hiring was compared to the appropriate population figures, we found the question of discrimination sufficiently doubtful to *40remand the case for a proper hearing. This union had a minority membership of 5.77 percent in December, 1976, fairly close to the figure in Local 14. As the district court found, less than one-third of the membership has joined this union since the effective date of Title VII. See Local 638, supra, 532 F.2d at 824; 401 F.Supp. at 474. Moreover, the record does not reveal how much control the union had over direct transfers and organization of non-union shops.3 Nor does the majority take account of the extremely depressed conditions in the construction industry in New York during the period in question. Finally, the district court did not have before it the “applicant-flow data” deemed essential by the Supreme Court in Hazelwood. See —U.S. at —, 97 S.Ct. 2736 (Brennan, J., concurring). Without clear findings on these matters, we are in a position to affirm neither the finding of discrimination nor the broad remedial order. As the Hazelwood Court concluded:
It is thus clear that a determination of the appropriate comparative figures in this case will depend upon further evaluation by the trial court. As this Court admonished in Teamsters, “statistics . come in infinite variety [TJheir usefulness depends on all of the surrounding facts and circumstances.” [431 U.S. at 340, 97 S.Ct. 1843]. Only the trial court is in a position to make the appropriate determination after further findings. And only after such a determination is made can a foundation be established for deciding whether or not Hazel-wood engaged in a pattern or practice of racial discrimination in its employment practices in violation of the law.
—U.S. at —, 97 S.Ct. at 2744 (footnote omitted).
I do not mean to suggest that this union’s history in racial matters is commendable. The blatant pre-Act discrimination is deplorable, the very most that can be said for it is that it was not unlawful. Furthermore, there have been enough subsequent acts of discrimination to support a conclusion that compliance with Title VII has been, at most, grudging and half-hearted.4 Nonetheless, Title VII does not empower us to right all the wrongs of society, but only to correct specific illegal conduct. Equal Employment Opportunity Commission v. Enterprise Ass’n Steamfitters Local 638, 542 F.2d 579, 592 (2d Cir. 1976), cert. denied, 430 U.S. 911, 97 S.Ct. 1186, 51 L.Ed.2d 588 (1977). It may well be that the plaintiffs in this action could prevail under a proper test. However, no matter how “just” the result appears, we cannot base a finding of liability on “a wholly inappropriate legal standard of discrimination,” Hazelwood, supra, — U.S. at —, 97 S.Ct. at 2744 (Brennan, J., concurring), as was done here. Thus, further factfinding is required, which can be accomplished only if the case is remanded to the district court. Id. (majority opinion).
III.
There is a second fundamental error in the determination of liability. The district court used the New York City minority population as the standard to determine the existence of discrimination. Under our recent decision in Local 14, this was clearly erroneous. As Judge Van Graafeiland stated for a unanimous panel:
[W]here a significant number of union members come from outside the union’s geographic jurisdiction, the court must widen its sights; the appropriate reference area then should be that region from which the union draws its members.
553 F.2d at 254. In June, the Supreme Court, in Hazelwood, “highlighted] the importance of the choice of the relevant labor market area.” — U.S. at — n.17, 97 *41S.Ct. at 2744.5 The majority ignores the district court’s error in this crucial decision.6
The district court below determined the minority population of New York City to be 29 percent, which it used as a benchmark. 401 F.Supp. at 488-89. This was clearly incorrect in light of Hazelwood and Local 14. The error is underscored by the fact that the court below expanded the areas from which minority membership would be drawn on the ground that:
If it be true that many of the present membership reside outside the limits of New York City there is no reason why applicants should be restricted to New York City.
421 F.Supp. 603, 618. It may be true, as the majority states-, that this is a careful finding of fact that an “insignificant” part of the membership lives outside of New York. I must note, however, that “insignificant” is generally not synonymous with “many.” If so, it belies both our earlier decision that the relevant labor pool was “the Metropolitan area,” 532 F.2d at 831, and the position taken on this appeal by the EEOC that “many of the present membership reside outside the limits of New York City.” Brief of the EEOC at 19. It is abundantly clear that a further hearing is necessary, at which proper statistical evidence could be taken.
IV.
The crucial nature of this evidence is illustrated by a comparison with the Local 14 case. Local 15,7 the union involved there, was another construction union whose jurisdiction is New York City. Just as here, there was ample evidence of blatant pre-Act discrimination, resistance to change and substantial individual acts of discrimination after the Act. However unpleasant such conduct is, it does not necessarily make out a prima facie “pattern and practice” case under Title VII. In that case, we indicated that the proper hiring goal was approximately 16.2 percent. Here, four out of every five members taken into the union have come from the apprenticeship program. Minority participation in that program rose from .37 percent in 1965 to 21.8 percent in 1967, fell to 9.77 percent in 1973, and rose to 14 percent in 1974. These figures cast substantial doubt on the finding that the apprenticeship program was discriminatory. Compare Local 14, supra, 553 F.2d at 254; Hazelwood, supra, — U.S. at —, 97 S.Ct. 2736. Moreover, the district court confusingly lumped together pre and post-Act discrimination in such areas as the organization of non-union shops, preventing meaningful analysis of the record. 401 F.Supp. at 485-86. There is thus some possibility that the union is not liable under Title VII. In any event, if liability is found, a proper remedy cannot be formulated until the scope of the violation is known. The majority’s holding makes this task impossible.
V.
Even if the majority’s affirmance of the finding of liability is correct, the case still should be remanded for the formulation of a proper remedy. If the district court fixes the proper figure for determining the existence of discrimination, then it logically follows that the same figure is the appropriate target in any affirmative action plan imposed. The rule in Title VII cases, as in all others, is that the remedy is to be fitted to the wrong. Thus, in Albemarle Paper Co. v. Moody, 422 U.S. 405, 95 S.Ct. 2362, 45 L.Ed.2d 280 (1975), the leading case on rem*42edies under Title VII, the Court explicitly adopted this familiar principle:
“The general rule is, that when a wrong has been done, and the law gives a remedy, the compensation shall be equal to the injury. The latter is the standard by which the former is to be measured. The injured party is to be placed, as near as may be, in the situation he would have occupied if the wrong had not been committed.” Wicker v. Hoppock [73 U.S. 94] 6 Wall. 94, 99 [18 L.Ed. 752] (1867).
Id. at 418-19, 95 S.Ct. at 2372. See Dayton Board of Education v. Brinkman, —U.S. —, —, 97 S.Ct. 2766, 53 L.Ed.2d 851 (1977); Milliken v. Bradley, 418 U.S. 717, 738, 744, 94 S.Ct. 3112, 41 L.Ed.2d 1069 (1974). We expressed the same views in Local 14, where we stated:
When a District Court finds that discriminatory practices on the part of a union or an employer have prejudiced minority workers, it should frame its relief with an eye toward remedying the wrong, see generally Albemarle Paper Co. v. Moody, 422 U.S. 405, 421 [95 S.Ct. 2362, 45 L.Ed.2d 280] (1975), and should interfere with the defendant’s operations no more than is necessary to accomplish this result.
553 F.2d at 256.
The majority simply ignores this principle. Instead, it holds, in effect, that the district court, once it has made a finding of discrimination, is free to select a hiring goal on an arbitrary basis, or no basis at all. There can be no other explanation for a holding that, even if 16.2 percent is the appropriate figure for determining the existence of discrimination, 29 percent is the hiring goal which must be set to remedy that discrimination. A simple example will demonstrate the absurdity of this position.
Assume that two unions, in related trades, have jurisdiction over a city whose minority population is 30 percent. The metropolitan area of which the city is a part has an overall minority population of 15 percent. Both unions’ membership live throughout the metropolitan area. On the effective date of the Act, neither union has any minority members.
The first union in this example admits 15 percent minority members after the Act becomes effective; the second union only 10 percent. Under Hazelwood and Local 14, the first union is in full compliance with Title VII. The second union, however, would be in violation of the Act since it falls short of the 15 percent goal. A minority membership goal of 30 percent within five years could be set by the district court, following the majority reasoning of this case.
After five years, if the second union has succeeded in raising its overall minority membership to 20 percent, .it would be in violation of the affirmative action plan imposed. At the same time, the first union, with its 15 percent post-Act minority membership, would be in full compliance with the law. I do not believe that the drafters of Title VII envisioned such an incongruous result.
VI.
Finally, this decision will, on balance, retard rather than advance the achievement of non-discrimination in employment. While it may lead to somewhat “more” integration in the central cities, it will halt efforts to integrate suburban employers. Thus, if a union whose jurisdiction is Nassau and Suffolk Counties were found liable under Title VII, I assume that the majority, in order to be consistent, would set the hiring goal by reference to the small minority population of those suburban counties. It would presumably be error to consider New York City’s far more substantial minority population. Although this result appears to be compelled by today’s decision, it is clear that it will hardly advance the cause of integration. If we are to eliminate the national disgrace of employment discrimination, we must widen, and not restrict, our horizons. I believe that the majority, proceeding from the best of motives, jeopar*43dizes the cause Title VII is designed to serve.8
I would vacate the decision of the district court, and remand for a hearing under proper standards.
. Of course, the Supreme Court is not bound by our law of the case. IB Moore’s Federal Practice fl 0.404[10], at 574 (2d ed. 1974). Since Hazelwood and Teamsters are decisions of the Supreme Court, our prior decisions to the contrary in this case will carry no weight on certio-rari, and applying law of the case is an exercise in futility. White v. Higgins, 116 F.2d 312, 317 (1st Cir. 1940) (Magruder, J.). The only effect can be further proceedings and waste of time, precisely the evils which law of the case is meant to avoid.
. In Wright, the conduct involved, organizing a “union shop,” had actually been unlawful at the time it was enjoined. Here, by contrast, the injunction was based upon an erroneous construction of Title VII. Also, Wright was a stronger case for applying law of the case, since it involved a consent decree rather than, as here, a fully litigated controversy.
Both Swiñ and Wright state that an injunction must be modified if transformed by a change of law into an “instrument of wrong.” However, Wright makes it clear that this means merely that a party is restrained from carrying out what has become a legal course of conduct. The opinion states:
Had the 1945 decree simply represented relief awarded by the District Court after a trial of the action instituted by petitioners, there could be little doubt but that, faced with the 1951 amendment to the Railway Labor Act, it would have been improvident for the court to continue in effect this provision of the injunction prohibiting a union shop agreement as being unlawful per se, or its use as an instrument to effectuate other statutorily forbidden discriminations. That provision was well enough under the earlier Railway Labor Act, but to continue it after the 1951 amendment would be to render protection in no way authorized by the needs of safeguarding statutory rights at the expense of a privilege denied and deniable to no other union.
364 U.S. at 648, 81 S.Ct. at 371. See Theriault v. Smith, 523 F.2d 601 (1st Cir. 1975) (vacation of consent decree to conform with subsequent decision of Supreme Court); 11 C. Wright & A. Miller, Federal Practice and Procedure §§ 2863, 2961 (1973); Developments in the Law — Injunctions, 78 Harv.L.Rev. 994, 1082 (1965).
. As is clear from the district court opinion, the discriminatory organization of non-union shops mentioned in the majority opinion at 36 n.8 took place in “the late 1950’s and early 1960’s,” 401 F.Supp. at 485, before the effective date of the Act.
. It is for the district court to determine if these constitute a “pattern and practice” of discrimination or if they call for individual treatment. Local 14, supra, 553 F.2d at 255-56.
. Hazelwood dealt with the exact opposite of this situation. There, a district court failed to include the minority population of a central city in a lawsuit against a largely white suburb. It is inconceivable, however, that a different legal standard determines the existence of discrimination in cities and suburbs.
. One commentator has noted that “the geographical delimitation of the relevant population can be the deciding factor in a case.” Lopatka, A 1977 Primer on the Federal Regulation of Employment Discrimination, 1977 U. Ill. L.F. 69, 76.
. Local 14 itself was found liable for discrimination. The remand was ordered for Local 15, a companion union in the same case.
. As we stated in Local 14:
We must carefully balance the need for effective enforcement of the Act against overzealous enforcement which can only lead to resentment and a resistance to change.
553 F.2d at 255.