dissenting.
In a carefully considered and well written opinion, District Judge Horace W. Gilmore concluded that the Census Bureau must adjust its final 1980 census totals to correct the widely acknowledged and differential undercounting of Hispanics and black Americans. Today, the majority reverses in an opinion that is rife with legal and factual misanalysis. I must respectfully dissent.
I.
Although the majority opinion contains a substantial amount of far ranging dicta, its holding is limited, technical, and I think, myopic.1 It rules that this case is not justi-*627dable because either the plaintiffs lack standing or because the controversy presented by the parties is not sufficiently ripe. Judge Merritt, speaking for the ma*628jority, rules that “[t]he plaintiffs have shown no judicially cognizable injury and lack standing to sue. The claimed injury is based on a state of affairs not yet in exist*629ence, and it is so hypothetical in nature that it does not present a controversy capable of judicial resolution.”
I respectfully disagree with both parts of this analysis. I believe that the court misuses standing and ripeness doctrines in order to avoid grappling with a serious question of constitutional harm.2 I would reach the merits, and although I am in substantial agreement with the district court’s analysis, I would remand this case to the district court for further factfinding in light of data which has become available subsequent to the district court’s decision.
A. Standing
I agree that the majority correctly outlines the modern law of standing. In Association of Data Processing Service Organizations v. Camp, 397 U.S. 150, 90 S.Ct. 827, 25 L.Ed.2d 184 (1970) and Barlow v. Collins, 397 U.S. 159, 90 S.Ct. 832, 25 L.Ed.2d 192 (1970), the Supreme Court established the injury-in-fact requirement for constitutional standing. Subsequent decisions have made clear that in order to show injury-in-fact, a plaintiff must establish a “distinct and palpable injury” and a fairly traceable causal connection between the claimed injury and the challenged conduct. See Gladstone Realtors v. Village of Bellwood, 441 U.S. 91, 99 S.Ct. 1601, 60 L.Ed.2d 66 (1979); Orr v. Orr, 440 U.S. 268, 271-73, 99 S.Ct. 1102, 1107-08, 59 L.Ed.2d 306 (1979); Duke Power Co. v. Carolina Environmental Study Group, 438 U.S. 59, 72-3, 98 S.Ct. 2620, 2629-30, 57 L.Ed.2d 595 (1978). A plaintiff can satisfy the causation requirement by showing that there is a “substantial likelihood” that the relief requested of the court will redress the claimed injury. Duke Power Co. v. Carolina Environmental Study Group, supra, at 75 n.20, 98 S.Ct. at 2631 n.20; J. Nowak, Handbook On Constitutional Law 73-9 (1978); id. 1978 Pocket Part at 7; L. Tribe, American Constitutional Law 92 (1978). Although the court correctly outlines the law of standing, it misapplies that law to the facts of this case.
The majority does not seriously dispute the fact that plaintiffs have established the requisite injury for standing purposes. The district court found below that the plaintiffs would lose a proportionate amount of political representation equal to the amount of differential undercounting in the federal census. There is little question that such a loss of political representation constitutes the kind of “distinct and palpable” injury necessary to establish constitutional standing. See, e. g., Kirkpatrick v. Preisler, 394 U.S. 526, 89 S.Ct. 1225, 22 L.Ed.2d 519 (1969); Wesberry v. Sanders, 376 U.S. 1, 84 S.Ct. 526, 11 L.Ed.2d 481 (1964); Baker v. Carr, 369 U.S. 186, 82 S.Ct. 691, 7 L.Ed.2d 663 (1962). But the majority concludes that the plaintiffs nevertheless lack standing because the second prong of the injury-in-fact requirement is not satisfied. It holds that the plaintiffs’ loss of representation is not caused by the Census Bureau’s differential undereounting of minorities, but is caused solely by the redistricting decisions of the Michigan state legislature. It bases this finding of intervening causation on the fact that the state legislature is not required by law to use the sub-state population totals supplied by the Bureau. If the state legislature chooses to use Bureau-supplied figures for state reapportionment, the majority holds that this decision would be an independent act breaking the chain of causation between the challenged actions of the Census Bureau and the injury to the plaintiffs. I cannot accept this conclusion because, in my view, it construes the causation component of the injury-in-fact requirement far too strictly.
The causation component of modern standing analysis has been grafted from the causation theory of the law of torts. In tort law, two forms of causation are recog*630nized. The first is “but for” causation or the sine qua non rule. The second is “proximate” or “legal” causation. I think it is clear that under either approach causation exists in the circumstances of this case for purposes of establishing constitutional standing.3
In “but for” causation, a defendant’s conduct is not a cause of the event if the event would have occurred without it. See W. Prosser, Law of Torts 238-9 (4th ed. 1971); Smith, Legal Causes in Actions of Tort, 25 Harv.L.Rev. 103-9 (1911). Under this formula, if injury would not have occurred to the plaintiff but for defendant’s conduct, the defendant is liable for the whole injury “regardless of its position in the string of acts leading to the injury though one or more of the other causes contributing to the result also involved wrongdoing on the part of other persons.” F. Harper & F. James, 2 The Law of Torts 1121-22 (1956).4
Proximate cause, on the other hand, refers to the “more or less undefined considerations which limit liability even where the fact of causation is clearly established.” Prosser, supra at 244. Under proximate causation theory, an act occurring after the defendant’s action and before the plaintiff’s injury can be an “intervening cause”, i. e. one which erases any responsibility on the defendant’s part. However, “[i]f the intervening cause is one which in ordinary experience is reasonably likely to be anticipated, or one which the defendant has reason to anticipate under the particular circumstances, he may be negligent only for that reason.” Id. at 272. In other words, “[f]ore-seeable intervening forces are within the scope of the original risk, and hence of the defendant’s negligence. The courts are quite generally agreed that intervening causes which fall fairly in this category will not supersede the defendant’s responsibility.” Id. at 273.5
*631If the Census Bureau failed to supply sub-state population figures, containing a differential undercounting of minorities, the state legislature could not then reapportion on the basis of those figures. Therefore, “but for” cause is present. As previously mentioned, the district court found that the plaintiffs would lose a proportionate amount of voting representation equal to the amount of the differential under-counting in the federal census. This is the claimed injury of the plaintiffs. It was found by the district court to exist, and for standing purposes this finding is entitled to great weight. The plaintiff could not suffer this injury which was found to exist by the district court “but for” the Bureau’s challenged action. Thus, the causation requirement under the first formulation is necessarily met.
The Census Bureau’s differential under-counting is also the proximate cause of the plaintiff’s injury. It is true that the under-count is not the direct cause of the state’s malapportionment.6 However the Michigan legislature’s redistricting decision is not an intervening cause since the legislature’s use of the defective figures is not an unforeseeable consequence of the Bureau’s actions. The parties do not dispute the fact that in its history Michigan has never failed to use federally supplied sub-state population figures. Moreover, the Bureau should know that in recent American history the states have almost invariably used federally-supplied figures for reapportionment.7 In addition, the Bureau should have been aware of the fact that if the state chooses not to use the federally-supplied figures it would have to use data at least as accurate as that supplied by the Bureau. See Burns v. Richardson, 384 U.S. 73, 86 S.Ct. 1286, 16 L.Ed.2d 376 (1966). Michigan has no already established bureaucracy to gather such data and perhaps has neither the financial nor technical capacity to do so. In fact, since 1959 state law has provided that the state and sub-state totals of the federal *632decennial census shall be used by the State of Michigan for all election purposes.8
For the reasons outlined above, it is certainly foreseeable that the state would reapportion on the basis of the inaccurate figures. Therefore, any loss of state representation suffered by the plaintiffs is within the scope of the foreseeable risk created by the Bureau’s failure to accurately count minorities. Accordingly, the Bureau’s action has to be a cause of the injury and plaintiffs have standing to challenge the conduct.
The majority finds that the decision of the state legislature to reapportion on the basis of federally-supplied totals breaks the chain of causation between plaintiffs and the Bureau because as “[a]n independent third party, the Michigan state legislature would play a necessary role in determining the effects of the census upon plaintiffs.” I think this rationale misses the point. The state legislature’s action would certainly be necessary in order to create the plaintiffs’ injury, but this fact does not mean that other concurrent causes are not present as well. If plaintiffs’ injury would not occur “but for” the Bureau’s differential under-counting, then in my view the undercount-ing is a concurrent cause of the injury and the causation requirement is satisfied.
However, if “but for” causation is not enough to establish standing here, then certainly proximate cause does. If the Bureau could reasonably foresee that the state legislature would use the figures in such a way as to cause the claimed injury to the plaintiffs, then the actions of the state — even if it is an independent party — cannot break the chain of causation.
Recent decisions support my view of the causation requirement’s application here. In Duke Power Co. v. Environmental Study Group, supra, the district court discerned a “but for” causal connection between the existence of the Price Anderson Act, the construction of nuclear power plants in the country, and the adverse consequences alleged by the plaintiffs. The petitioners challenged this finding of causation on two grounds in the Supreme Court: (1) that the particular nuclear power plants in question would have been constructed by private companies without the Price Anderson Act’s limitation on liability; and (2) that if Congress had not passed the Price Anderson Act it may well have chosen to pursue the nuclear program as a government monopoly — as it had done from 1946 to 1954. The Supreme Court rejected both contentions. It held that industry testimony at Congressional hearings on the passage of the Act indicated that the industry would not develop nuclear power facilities without a limitation on liability. In addition, it rejected the contention that the Price Anderson Act was not a “but for” cause of any adverse consequences to the plaintiffs since the government would have undertaken development of nuclear power on its own. It held that “[wjhatever the ultimate accuracy of this speculation, it is not a response to the simple proposition that private power companies now do in fact operate the nuclear-powered generating plants injuring the ap-pellees, and that their participation would not have occurred but for the enactment and implementation of the Price Anderson Act.” Id., 438 U.S. at 77-8, 98 S.Ct. at 2632-33. (emphasis added)
In Duke Power, “but for” causation was met because the claimed injury of the plaintiff could not have occurred without the presence of the Price Anderson Act. In this case, the plaintiffs’ injury would not occur without the defendants’ differential under-counting of minorities. Therefore, causation for purposes of standing has to be present in the instant case as well.
In Duke Power, the direct cause of the plaintiff’s injury was the decision of the private utility companies to construct the nuclear power plants at issue. The effect of the Price Anderson Act was one step *633removed. Nevertheless the Supreme Court found a causal nexus between the operation of the Price Anderson Act and the claimed injury of the plaintiffs. As I have previously mentioned, an act does not have to be the last in a “string of acts leading to the [plaintiff’s] injury” in order for it to be a “but for” cause. See F. Harper and F. James, supra at 11. The sole issue is whether the injury would have resulted without the occurrence of the act. In this case, the defendants’ challenged action is not the direct cause of plaintiffs’ injury. However, as in Duke Power, the plaintiffs’ injury would not result without the occurrence of the defendants’ challenged conduct. Thus, causation sufficient enough to establish standing is present even though the defendants’ action is not a direct cause.
The Court in Duke Power did not specifically address the issue of intervening causation. But, it is clear that construction of nuclear plants and the resultant injury to the plaintiffs were foreseeable results of passage of the Price Anderson Act. This is evident because the Court did not find that the decision of the companies to construct nuclear power plants was an intervening cause that broke the chain of causation between the Price Anderson Act and the claimed injury. The company argued that in the absence of the Act, the government might have constructed such plants on its own or that private companies might have raised the necessary funding under an alternate scheme. However, the Court ruled that prior cases did not require “a party seeking to invoke federal jurisdiction to negate the kind of speculative and hypothetical possibilities suggested in order to demonstrate the likely effectiveness of judicial relief.” Id. at 78, 98 S.Ct. at 2633. The same analysis applies here. In order to demonstrate a traceable link from the Bureau’s actions to their claimed injury, the plaintiffs should not be required to negate the “speculative and hypothetical possibilities” that Michigan might: (1) change its long standing law requiring the sole use of federal census totals; and either (2) develop the capacity to conduct an accurate state census or (3) develop the capacity to make accurate adjustments of federal census data. The decision of the Michigan legislature to use raw federally-supplied sub-state totals is a foreseeable consequence of the Bureau’s supplying of those figures. Thus, as in Duke Power, there is no intervening causation present.
The case at bar is distinguishable from Simon v. Eastern Ky. Welfare Rights Organization, 426 U.S. 26, 96 S.Ct. 1917, 48 L.Ed.2d 450 (1976). In that decision, the Supreme Court found that § 501(c)(3) of the Internal Revenue Code was not the cause of the plaintiffs’ failure to receive free emergency medical treatment from the defendant hospital. The Court noted that: “It is purely speculative whether the denials of service specified in the complaint fairly can be traced to petitioners’ ‘encouragement’ or instead result from decisions made by the hospitals without regard to the tax implications.” Id. at 43, 96 S.Ct. at 1926. In other words, § 501(c)(3) was not a “but for” cause of the claimed injury and plaintiffs necessarily had no standing to bring the action. As previously discussed, that is not the case here; differential undercounting is a cause of the plaintiffs’ injury.
Simon apparently provides the basis for the majority’s invocation of intervening cause analysis in this case. In Simon, the Supreme Court did state that “Article III still requires that a federal court act only to redress injury that fairly can be traced to the challenged action of the defendant, and not injury that resuits from the independent action of some third party not before the court.” 426 U.S. at 41-2, 96 S.Ct. at 1925-26. (emphasis added). However, since § 501(c)(3) was not even a “but for” cause of the plaintiff’s alleged injury, the Court did not have to explain when for standing purposes an injury is caused by one precedent and not by another. It is extremely doubtful that the above language in Simon precludes concurrent causation in the law of standing.9
*634I think there can be no dispute about the fact that the Census Bureau’s differential undercounting is a concurrent cause here. Accordingly, for the reasons discussed above, I would find standing for the plaintiffs to bring this action.
B. Ripeness
The majority correctly outlines the Ripeness doctrine. Unfortunately it misapplies this doctrine, just as it misapplies standing. The majority holds that it would be better to wait until the Michigan legislature acts on reapportionment because only then would the issues here be particularized enough for judicial resolution. I disagree. I think the court’s invocation of ripeness as a reason to reverse the district court amounts to another smokescreen to justify ducking important questions. The issues presented by this case should be decided now.
First, for reasons outlined above, I cannot agree with the majority’s statement that “[hjeightened public sensitivity to the problems of census undercounts makes past reliance on census figures an uncertain predi-cator of future legislative action.” The court speculates that the Michigan legislature will depart from its invariable past practice and either adjust federally-supplied census data or use their own data for reapportionment. For the reasons outlined above, I see no reason to believe this will occur.
Second, there is no reason to defer to the judgment of a legislative body in this case. The key questions pertaining to the merits of this case are: Is there a differential undercount? Can it be corrected? To answer these questions requires the analysis of expert testimony. If the expert testimony shows that there is an ascertainable differential undercount which can be corrected, there is but one permissible conclusion— it must be corrected. A state legislature has no more expertise than a court does in this area, and it is neither necessary nor proper to refer this case to the Michigan legislature so that it can give us its advisory opinion on these questions.
In fact, I think that the Michigan legislature is not the proper forum for the issues presented by this case. A proposal in the Michigan legislature to correct census figures would provoke a heated controversy, in which the citizens of Detroit and the citizens of other areas suffering high differential undercounting would be politically handicapped. Their strength and representation in the legislature is directly weakened by the very phenomenon at issue — differential census undercounting. The net result of legislative consideration of this issue is likely to be a decision dictated by politics and not dictated by objective appraisal of the expert demographic testimony. It would probably be too much to expect state representatives from areas which have benefited politically from differential undercounting to in effect vote themselves out of office and increase the representation of the citizens of Detroit. The Supreme Court recognized this problem when it held in Baker v. Carr, 369 U.S. 186, 82 S.Ct. 691, 7 L.Ed.2d 663 (1962), that the protection of the Constitutional right of equal representation could not be left solely to the legislatures. Because the vindication of this most important of Constitutional rights is a responsibility of the courts, I think we cannot put off a decision on the merits here.
Third, and finally, I think that failure to rule now will cause hardship to both parties and to the public interest. The issues presented by this case are too important. Avoiding these issues as the majority does will not make them go away. At the most, all that the majority opinion does is temporarily change the forum — from federal court to the Michigan legislature. The Michigan legislature will then have to de*635cide whether to adjust the census figures and, if so, by how much. Regardless of what decision the legislature reaches, someone will be unhappy with it. We can be assured that if the Michigan legislature fails to adjust the census figures, or adjusts the figures only slightly, that the plaintiffs will then return to federal court. But the majority is in error in its assumption that the litigation will end if the Michigan legislature grants relief to the plaintiffs and adjusts the census figures in a manner acceptable to them. If the Michigan legislature does make an undercount adjustment, it is certain that disgruntled voters and/or elected officials will promptly run to federal court and file a challenge to the adjustment. The stakes involved are high, and any adjustment must meet the standards outlined in Wesberry v. Sanders, supra, and its progeny. As a result, there is every incentive for affected persons, i. e. those residing in areas containing relatively few minorities, to challenge any adjustment.10
Thus it is clear that the issues presented by this case will require judicial resolution sooner or later. The majority prefers later. Unfortunately, this will cause great hardship to all concerned. If a federal court disagrees with the state’s reapportionment determinations, the court will have to order reapportionment using the census figures it has determined are appropriate. The time, effort and money involved will be greater at that later date than now.
Given the political considerations discussed above, I think there is a significant risk that a court’s undercount determination will differ from the Michigan legislature’s. As I have said, if this happens there will be hardship to all concerned. Moreover, there will exist tremendous and unacceptable uncertainty regarding electoral boundaries until the undercount issue is resolved.
In sum, the issues presented are ripe and should be resolved now. It is unlikely that the Michigan legislature will change things or that its deliberations will be helpful to us. Most important, no matter what the Michigan legislature does, we will have to decide the issues presented by this case anyway. There is simply no reason to delay a decision on the merits. On the contrary, I think that it is the height of folly not to decide this case given both the importance of the issues and the problems engendered by postponing a decision.
II.
As previously stated, the district court below ruled that the defendants must adjust their national, state, and sub-state census totals for 1980 to correct the differential undercounting of blacks and Hispanics. The court also ruled that the defendants could select at their discretion any available methods of adjustment “so long as they [the methods] are statistically defensible.” These rulings are based on sound legal principles and are supported by the evidence at trial. However, new data released since the entry of judgment against the defendants leads me to the conclusion that the case should be remanded to the district court for further fact-finding, since this information was not available to the court at the time it reached its decision.
A.
I think that the district court’s formulation of the law governing this controversy *636is unassailable. Simply stated, the court held that the Secretary of Commerce and the Bureau of the Census are bound by the Constitution to conduct the decennial census as accurately as possible. This holding follows from the Supreme Court’s recognition in Baker v. Carr, 369 U.S. 186, 82 S.Ct. 691, 7 L.Ed.2d 663 (1962), of the constitutional right of equal representation — that one citizen’s vote should not weigh any more or less heavily than another citizen’s vote. Later, the Supreme Court went on to hold that equal representation requires that the apportionment of representation in the United States House of Representatives and in the state legislatures must be based “as nearly as is practicable” on precise mathematical equality. See Wesberry v. Sanders, 376 U.S. 1, 84 S.Ct. 526, 11 L.Ed.2d 481 (1964) and Kirkpatrick v. Preisler, 394 U.S. 526, 89 S.Ct. 1225, 22 L.Ed.2d 519 (1969).
It goes without saying that unless the population base for reapportionment is accurately determined, there can be no such precise mathematical equality between voting districts. If an inaccurate enumeration takes place, then some voting districts by virtue of differential undercounting would contain more persons than others.11 In these circumstances, the deleterious effect on equality of representation might be as serious as any that would occur from gerrymandering. See Note, Demography and Distrust: Constitutional Issues and the Federal Census, 94 Harv.L.Rev. 841, 860 (1981).
Since the sole constitutional function of the decennial census is to provide a population base for reapportionment of Congressional districts,12 the Census Bureau is bound by the Constitution to count people as accurately as possible.13 Otherwise, the promise of Baker v. Carr would, in my opinion, be rendered a teasing illusion.
B.
The district court found that the Secretary of Commerce and the Census Bureau did not conduct the 1980 census as accurately as possible since they did not adjust the final tally to correct for the differential undercounting of blacks and Hispanics. The court found persuasive the plaintiffs’ expert testimony which established that there has been a perennial differential un-dercounting of blacks and Hispanics. In fact, testimony by Dr. Philip M. Hauser, University of Chicago Professor of Urban Sociology and former Acting Director of *637the U. S. Census Bureau, established that under the original Constitution “the slaves [in 1790 were] . .. counted as three-fifths of a person, but [now almost 200 years later] the black male twenty-eight years of age counts as 78 per cent of a person.”14 The district court also found that the extent and nature of this differential undercount-ing could be accurately determined,15 and that the state of the statistical art allowed the Census Bureau to impute into the final census tally the number of persons missed *638during the headcount.16 According to the testimony, statistical adjustment would yield a more accurate estimate of the number of persons within the United States than would result from current Bureau counting procedures.17
Because the trial occurred before release of preliminary 1980 census data, the expert *639witnesses below based their testimony on data derived from past censuses and from projections of 1980 census results.18 The parties did stipulate that there would be an undercounting of persons within the United States in 1980, and that the Bureau would undercount blacks and Hispanics at a higher rate than whites.19 However, there was no stipulation as to the rate of this differential undercounting for 1980.
*640The only way to determine actual differential undercounts is to compare the census headcount with demographic estimates of the nation’s true population. As the majority notes, the Census Bureau argues on appeal that it cannot determine the actual differential undercount for 1980 because of the presence of large numbers of illegal aliens within the United States. Preliminary 1980 census figures released after the trial below show that the Bureau has counted approximately 226 million people. Since earlier demographic studies indicated that about 226 million people would be within the United States on April 1, 1980, the preliminary census data on the 1980 headcount substantially equals the available demographic estimates.20 Thus, the defendants assert that no statistical adjustment of the 1980 headcount is possible since there are no available means to measure how many illegal aliens have been counted. They conclude that there is no way to determine how many blacks and Hispanics were missed in the 1980 headcount.
In its recitation of the facts, the majority seems to accept the Bureau’s analysis as true.21 However, the plaintiffs have not had an opportunity to rebut the Census Bureau’s analysis since the plaintiffs’ experts testified before the release of the preliminary 1980 census data. It is for this reason that I would remand this case to the district court for further fact-finding. If the plaintiffs can show that there are available methods to measure the differential undercounting of blacks and Hispanics, then the district court’s conclusion is correct — statistical adjustment to correct the differential undercounting is feasible. I think the plaintiffs should have this opportunity. They should be allowed to submit additional evidence on remand, particularly in view of the evidence introduced in Carey v. Klutznick, 508 F.Supp. 420 (S.D.N.Y. 1980) which indicates that there may be ways to accurately measure the 1980 under-count.22
*641Accordingly, I would remand this case to the district court.
. In his opinion for the majority, Judge Merritt makes a number of statements in dicta which have no support in the record. Specifically, the majority states:
No claim is made that upper and middle class blacks and hispanics are undercounted in comparison with whites of the same economic class. Rather, plaintiffs have demonstrated the undeniable fact that for a complex set of reasons it is more difficult to count the poor and the uneducated whether black or white, than the middle class and the well-to-do. Plaintiffs rely on Census Bureau studies of previous census figures that show that as a group, more blacks are missed than whites because as a group, blacks are more disadvantaged. Plaintiffs’ key witness, Dr. Philip M. Hauser, a former director of the Census Bureau and chairman of the University of Chicago Sociology Department, stated the basic reason for previous undercounts of the black population: “It is not a question of color; it is not the difference in color that makes them hard to count. It is that whole cluster of characteristics associated with col- or which stands as a proxy for this cluster of characteristics.” (citations omitted)
*627Proceeding from this premise, the majority goes on to conclude that a synthetic adjustment technique might be unfair:
The major portion of an estimated 1,000,-000 black undercount (approximately 4%) apparently comes from the approximately 30% of the black population at or below the poverty line because they are hard to count. Yet the synthetic method would add 70% of the black undercount to the voting strength of the more advantaged group, the black middle class. As in the case of whites, distribution of the black undercount according to the synthetic method has the effect of taking voting strength from the disadvantaged black population and giving it disproportionately to the more advantaged members of the same class.
The majority opinion presents an amusing sociological analysis. It would be nice if the court’s views had some support in the record. Unfortunately, they do not.
All of the expert witnesses agreed that there is an undercount and that no census can count everyone. As the tables reproduced in the majority opinion show, the undercount varies according to three major variables: age, sex and race. In terms of electoral voting power, the significant variable is, of course, race. Dr. Carl Taeuber’s undisputed testimony below established that the black community of this country is concentrated in urban areas. Thus, a disproportionate racial undercount, combined with geographical concentration of blacks, results in underrepresentation of those areas with concentrated black populations.
As noted above, the majority claims that the racial variable in census undercounts is a myth. According to the majority, the racial variable is really an economic variable; it is poor people who are undercounted. The only reason blacks are undercounted more than whites, according to the majority, is that blacks are poorer than whites.
The record contradicts these bald assertions. Dr. Hauser did state that the lower socioeconomic status of blacks was “the most important factor in the undercount that is measured by race.” (App. at 421). However, he stated that the undercount was “due to a number of factors.” (App. at 402). Dr. Hauser’s testimony was well summarized in Judge Gilmore’s opinion below:
Extensive expert testimony as to the underlying causes of the differential undercount was taken. Although the precise weight to be attributed to the various factors involved is not known, all of the experts agreed that some or all of the cluster characteristics associated with race are major factors. In this regard, Dr. Philip M. Hauser, a witness for the plaintiffs, stated that the following characteristics should be considered important:
1. lower income leading to diverse and unique housing patterns and living arrangements which frequently involve the lawful or unlawful conversion of housing units to accommodate increased numbers of people;
2. greater antagonism or resistance to government;
3. relatively less education; and
4. increased numbers of individuals within the harder-to-enumerate categories, such as young males.
497 F.Supp. at 1327-28
Examination of Dr. Hauser’s testimony reveals that race is indeed the critical factor in gauging an undercount. All that Dr. Hauser is saying in the passage quoted by the majority is that the combination of the factors of poverty, alienation, less education, and youth correlate with race and it is these factors which explain why there is an identifiable racially differential un-dercount of blacks and Hispanics. This is clear upon examination in context of Dr. Hauser’s testimony quoted by the majority:
Remember this, a third of all blacks live in poverty by the government standards. Twenty-seven per cent of Hispanics live in poverty by government standards. The combination of their socioeconomic status, residences which are often of substandard quality, hard to locate, residential patterns of occupancy which make it very difficult to find separate households within cramped quarters, converted structures, some legally converted, some illegally converted, that the socioeconomic status of our minority groups is such that by reason of that status and other research by the Bureau tends to support this, that they are difficult to count. It is not certainly a question of color, it is not the difference in color that makes them hard to count. It is that whole cluster of characteristics associated with color which stand as a proxy for this cluster of characteristics. Are these characteristics that result from a history of prior discrimination against minority groups in this country? Well, that is certainly clear, yes.
App. at 404-05
In fact, Dr. Hauser’s position became crystal clear in later testimony:
[Cjertainly no one is going to argue that the reason for the differential and undercount is traceable by the color of the skin or the ethnic origin. The reason for the undercount is traceable to the whole cluster of characteristics of these people for whom the designation of race or origin is a proxy. Now it turns out, as census statistics make quite clear, that these two populations are socially and economically greatly disadvantaged when compared with whites. They have lower education, they have much lower in*628come, they live in much larger proportions of substandard dwellings, there are much larger proportions of people living in poverty, as measured by government standards itself. They are highly concentrated and segregated within the inner cities of metropolitan United States as my colleague, Dr. Taeuber, will be indicating in greater detail later. Now these clusters of characteristics that characterize respectively the black and the Hispanic populations are pretty uniformly distributed through the black and Hispanic populations in the areas in which they are predominantly concentrated, the inner cities of our large metropolitan areas. And so, if you make these adjustments, you are very definitely I think correcting what amounts to be the most grievous aspects of census undercount.
App. at 432-33 (emphasis added).
Additional support comes from Dr. Carl Taeu-ber’s testimony:
The estimate of a substantial black under-count has been made for each census, going back to this period when blacks were very heavily rural and southern. As blacks have moved to the cities and to the North, this differential undercount has persisted. This is a kind of historical suggestion that it’s not simply because the blacks were rural or because Hispanics were migrant laborers. The undercount persists even as these populations have become heavily metropolitan. Now in fact, for the recent censuses, particularly for 1970 and from what we know already of what the 1980 results will show, we know that the blacks and Hispanics are still subject to a considerable amount of housing discrimination and job discrimination, which tends to help produce these concentrations, the residential segregation, which means that the middle class members of these groups as well as the working class, the poorer members, all tend to live predominantly in these ghettos or barrios. They are not widely dispersed by social class in the same degree that the white population is. We know that within these concentrated minority areas there is a very low social and economic status. There are many broken families, there is a very high rate of residential mobility, people moving from one rental unit to another, there are many unusual household and living arrangements, relatives, boarders, and the like. There is within these populations a fairly high degree of antagonism toward government, as represented in some of the riots we have recently had, as represented in a lot of the concerns say in the Hispanic community about problems of deportation, concern with the legal status of the people in relation to the government, fear of cooperation with the census and other government agencies. Now this would suggest that we should get — that the circumstances are generally similar from one ghetto to another, that the general features which may produce the undercount, as I must say we don’t know exactly what those features are, we do know that many blacks and many Hispanics are missed, but we don’t know exactly why. Now something in this cluster of attributes may be pertinent, and I think it has a lot to do simply with minority status rather than simply living arrangements because the Bureau, not just in 1980, the Bureau of the Census also for 1970, for which we do have major undercount status, did make major efforts to enumerate, did have special campaigns and the like, was aware of the problem ahead of the census, did the best they could and still came up with a substantial differential undercount. Now if every one of those ghettos and barrios is expected to have a substantial undercount, then the application of the synthetic method fits the criteria that there is a high degree of similarity within each of the concentrations, so that assuming that the national rate applies to each one of these areas, is a reasonable assumption, it is a very plausible assumption. Q. Now let us assume that we know that at the national level then blacks are undercounted at a rate of 8 per cent and whites at a rate of 2 per cent. Let us also assume that Hispanics are undercounted at a rate of 8 per cent compared to non-Hispanics at a rate of 2 per cent, the same four to one differential. In your opinion then, would an adjustment of the black and Hispanic population of each state and sub-state area by 8 per cent and an adjustment of the white non-Hispanic population by 2 per cent get us closer to the truth than the raw unadjusted census data?
A Yes, it would.
App. at 623-27 (emphasis added).
The majority baldly asserts that “no claim is made that upper and middle class blacks and hispanics are undercounted in comparison with whites of the same economic class.” Dr. Taeu-ber’s testimony refutes this assertion:
Q And was it the thrust of your testimony that from a sociological and demographic perspective, there is no reason to believe that the undercount rate of blacks and Hispanics varies significantly from one place to another?
A Yes, that was the thrust of it ...
The above testimony refutes the majority’s simplistic suggestion that undercounts are class-based and not race-based — that the few blacks or Hispanics in the suburbs are under-counted to a significantly lower degree than their counterparts in other areas. In fact, no authority that I have seen or heard of makes that suggestion. Under these circumstances, I can only wonder whether the author of the majority opinion read the same record that I did.
. See Note, The Supreme Court, 1975 Term, 90 Harv.L.Rev. 54, 212 (1976); L. Tribe, American Constitutional Law 93 (1978). See also Lewis, Constitutional Rights and the Misuse of Standing, 14 Stan.L.Rev. 433, 439 n.22 (1962) (relying upon standing doctrine “may lead to a shorter, more easily written opinion, but it will also lead to inadequate consideration of the real issue or conceal the actual thought process which produced the result.”)
. Causation analysis exists in areas of the law other than torts and standing. Although the analysis may be phrased in different ways when it appears in different areas of the law, upon closer examination causation can only be viewed in terms of either “but for” cause or proximate cause wherever it appears. For example, causation in the substantive criminal law and in criminal procedure takes the form of either “but for” or proximate cause depending on the circumstances. See 7 Wharton’s Criminal Law § 193-211 (1974) (discussing the causal connection between offender and offense). See also United States v. Crews, 445 U.S. 463, 100 S.Ct. 1244, 63 L.Ed.2d 537 (1980); United States v. Ceccolini, 435 U.S. 268, 98 S.Ct. 1054, 55 L.Ed.2d 268 (1978); Wong Sun v. United States, 371 U.S. 471, 83 S.Ct. 407, 9 L.Ed.2d 441 (1963) (cases applying the “fruit of the poisonous tree doctrine” under the Fourth Amendment’s exclusionary rule).
Since these two principles — proximate and “but for” cause — represent the only available modes of causation analysis in the law, it is reasonable to expect the Supreme Court to adopt them as it provides further explication of the second prong of the injury-in-fact constitutional standing requirement. Since it would be unreasonable to expect the Court to invent new causation principles which would operate solely in the law of standing, the tort causation analogies relied upon in this dissent should control here.
. A “refined” variation of “but for” causation is the substantial factor theory. Under this approach, a defendant’s conduct is a cause of an event if it is “a material element and a substantial factor in bringing it about.” Pros-ser, supra at 240. However, even under this variation of “but for” causation a defendant “will not be absolved from liability merely because other causes have contributed to the result, since such causes, innumerable, are always present. In particular, however, a defendant is not necessarily relieved of liability because the negligence of another person is also a contributing cause and that person, too, is to be held liable.” Id. at 240-41.
. The Supreme Court has not specifically decided which of these forms of causation analysis governs the second prong of the injury-in-fact standing requirement. See Note, The Causal Nexus: What Must be Shown to Sue in Federal Courts, 29 Univ. Florida L.Rev. 250, 256 (1977). It is clear, however, that a causation requirement for standing purposes which is more strict than that required in the tort law is impermissible. See note 3 supra. In fact,
“unlike tort cases, standing questions do not involve the liability of a private defendant for the consequences of his actions, but instead are concerned with the individual’s rights to obtain a judicial forum to protect himself from injurious government action. Since it is a threshold determination, the standing test should not be as strict as that which must be met on the merits. Instead the courts should be satisfied with a demonstration of a reasonable possibility of a causal relationship. Id. at 257-8 (footnotes omitted) (emphasis added).
*631The Supreme Court has implicitly recognized this principle. In Duke Power Co. v. Carolina Environmental Study Group, supra, 438 U.S. at 75 n.20, 98 S.Ct. at 2631 n.20, the Court expressed the “but for” causation standard for standing purposes in the following terms: Whether “there is a ‘substantial likelihood’ that the relief requested will redress the injury claimed.” (citations omitted). In other words, if there is a “substantial likelihood” that an injury would not occur without the challenged conduct, standing causation exists. See id. at 74, 98 S.Ct. at 2630. This formulation is less strict than the tort “but for” cause standard. In torts, substantial likelihood will not suffice; a plaintiff must show that its injury absolutely would not occur “but for” the allegedly tortious conduct.
I think that a showing of “a reasonable possibility of a causal relationship” and more was made below. Therefore, I believe that the second prong of the injury-in-fact requirement was met, and that the plaintiffs have the required standing to bring this action.
. The direct cause or most “immediate cause” of the plaintiffs’ claimed injury is concededly the decision of the Michigan legislature to use unadjusted federal census figures for state reapportionment. But this does not mean that the Bureau’s actions are not also a cause of the injury. Although Lord Bacon stated that the common law “contenteth itself with the immediate cause; and judgeth of acts by that, without looking to any further degree”, Smith Legal Cause in Actions of Tort, 25 Harv.L.Rev. 103, 106 (1911) that has long ceased being the case. See Prosser, supra at 246.
The decision of the Michigan legislature to use federal census figures is an act which occurs between the time that the Bureau supplies unadjusted figures and the time of plaintiffs’ injury. Therefore, the Bureau’s act is an indirect cause. But under tort law this alone would not eliminate the Bureau’s actions as a cause. If this is not the result in tort law then surely it cannot be the result in the law of standing. See note 4 supra.
. The government points out that Hawaii apparently does not use federally supplied sub-state totals for state reapportionment. See Reply Brief For the Appellants at 5 n.7 (citing Congressional Districts in the 1970’s, Congressional Quarterly 1973 at 51). However, there is no question that this is a very unusual state practice. See Note, Demography and District: Constitutional Issues of the Federal Census, 94 Harv.L.Rev. 841, 861 (1981) (“in practice, most states will probably use unadjusted figures, if those are the data supplied by the Census Bureau.”). It is so unusual that the Bureau must expect and be able to foresee that in almost every instance its sub-state totals will in fact be used by the states for reapportionment.
. M.S.A. § 2.212(22), M.C.L.A. § 8.3v reads as follows:
The population of the state or any political division thereof shall be determined, unless otherwise specifically provided, on the basis of the latest federal decennial census preceding the time as of which the population is to be determined.
. See Note, The Supreme Court, 1975 Term, supra, 90 Harv.L.Rev. at 209 n.31 (“The Court states that a federal court cannot act to ‘redress . .. injury that results from the indepen*634dent action of some third party . . To the extent that this implies that plaintiffs must prove that defendants’ actions were the sole cause of plaintiffs’ injury, it is certainly an incorrect view of causation as an element of a cause of action.”) (citations omitted)
Should the Court decide to adopt an intervening cause formula for standing, it is reasonable to expect that this formula will conform with pre-existing notions of intervening cause. See text accompanying note 5.
. The situation here is analogous to that presented by a voluntary affirmative action plan. Before the Supreme Court’s decision in United Steelworkers v. Weber, 443 U.S. 193, 99 S.Ct. 2721, 61 L.Ed.2d 480 (1979), an employer faced an impossible dilemma if it considered an affirmative action plan to remedy its possible past discrimination. If it failed to remedy past discrimination, it might face lawsuits from minority employees who might have been discriminated against. However, if the company initiated an affirmative action plan, it might face lawsuits from white workers. See Weber v. Kaiser Aluminum Co., 563 F.2d 216, 227-39 (5th Cir. 1977). (Wisdom, J., dissenting). In order to help solve the employer’s dilemma, the Supreme Court in Weber established a zone of reasonableness. See DPOA v. Young, 608 F.2d 671 (6th Cir. 1979) cert. pending.
Unfortunately, there is no Weber decision in the reapportionment area. As employers have faced discrimination and “reverse discrimination” lawsuits, so state legislatures will face reapportionment and “reverse reapporlionment” lawsuits.
. If the Bureau undercounted segments of the population in an equal fashion, then census undercounting would create no equal representation problem. Each voting district would be undercounted to the same degree, and the relative strength of each citizen’s vote would be unaffected by the undercount. The differential undercounting alleged here, on the other hand, dilutes the strength of those citizens who reside within voting districts with higher numbers of blacks and Hispanics since the elected representatives of these districts actually represent more people than do the elected officials of other districts.
. Article I, Section 2, Clause 3 of the Constitution provides:
Representatives .. . shall be apportioned among the several States ... according to their respective numbers ... The actual Enumeration shall be made within three Years after the first Meeting of the Congress of the United States, and within every subsequent Term of ten Years, in such Manner as they shall by Law direct.
. The Constitution requires the Secretary of Commerce and the Census Bureau to supply accurate national and state census figures. It does not require the Secretary or the Bureau to supply sub-state totals to the states for reapportionment of Congressional districts within the state. However, 13 U.S.C. § 141(c) requires the Bureau to do so. Since national and state totals are compiled by adding up sub-state census totals, it is obvious that the state or national census figures can be no more accurate than the local figures. Thus, the Bureau is necessarily bound by the Constitution to make its sub-state population totals as accurate as possible. It is bound by the statute to distribute these “accurate” figures to the states.
In addition, since “one man, one vote” is the law of the land, an accurate population base for reapportionment is necessary. The Census Bureau in virtually every case provides the data base for reapportionment. Thus, I believe that Baker v. Carr, 369 U.S. 186, 82 S.Ct. 691, 7 L.Ed.2d 663 (1962), and its progeny provide additional direct support for the Bureau’s duty to provide accurate census figures.
. As the majority notes, the Bureau estimates that it undercounted 7.7% of the blacks in 1970 and only 1.9% of the whites. But significant differential undercounting of black people has occurred throughout the history of the federal census. See R. Hill, The Synthetic Method: Its Feasibility for Deriving the Census Undercount for States and Local Areas (paper prepared for the Conference on Census Undercount, Feb. 25-6, 1980) reprinted in Appendix, at 1628. In 1870, 9.5% of blacks were uncounted, while 1.9% of the white population was not counted. Estimates for that year indicated that 41.9% of all the persons uncounted that year were black. Id. in 1920, 1930, and 1940 blacks were under-counted at the rate of 15.2%, 12.5% and 12.7% respectively. The white undercount, on the other hand, ranged from about one to less than three percent during those years. H. Alterman, Counting People: The Census In History 279 (1969). In 1950 and 1960 the black undercount was much greater than the undercounting of whites. Id. at 80-81.
Hispanics have experienced a similar history of differential undercounting in the federal census. Exact figures on the nature and extent of this undercounting in past censuses are not available, but the government does not dispute the fact that significant undercounting has occurred. The district court found that “almost as large a percentage of Hispanics have been undercounted” as blacks. 497 F.Supp. at 1318.
The expert testimony below established that a long history of faulty census questionnaires, language difficulties, as well as poverty, antagonism, lower education, and demographics have led to significant differential undercount-ing in the Hispanic community. See note 1 supra.
. The plaintiffs’ expert witnesses testified that because the “cluster of characteristics” in the nation’s black community has not dramatically changed since 1970 the rate that blacks are differentially undercounted should not change much, either. On direct examination, Dr. Hau-ser testified as follows:
Q. Dr. Hauser, if in 1970 we had an un-dercount of 2.5 per cent, then why do we have an undercount of this magnitude?
A. I think the answer is complex. It is due to a number of factors, including the following. One, we have a very large nation with over 3,000,000 square miles with very uneven distribution of population by a region, difficult to find. Second, and probably even more important, we have had tremendous differences in the social and economic status of our citizens and especially members of minority groups. This would include the blacks, Hispanics, and other groups. Blacks and Hispanics are the largest of this minority group. Third, we have had the factor of leadership, public opinion makers, including Congressmen, Senators, candidates for the President, all of whom generate attitudes that affect the citizen in his relationship to his government. And fourth, you have the citizen himself in a country such as ours where in an unparalleled fashion every citizen has the right to tell his government to go to hell, and many exercise that right, and particularly under the influence of a leadership which encourages this attitude. Now it is the interaction of all these factors, plus one other, to which I have referred in a specific instance. The changing composition of the population. If there is an increase in the proportion of the population that is most difficult to count completely, then there is an increase in the undercount.
Q. Is that situation occurring for 1980?
A. In 1980 these factors are all operative. The interaction and net effect of these is still to be determined. There are a number of factors that, all other things being equal, would tend to decrease the undercount, but there are also a number of factors operating which would tend to decrease [sic] the under-count and the net effect of this is not known to anybody at this time. App. 402-3.
Regarding the Hispanic population, Dr. Hauser testified as follows:
Q. You have explained how, through matching studies, it would be possible to obtain an estimation of the Hispanic under-count at the national level?
A. Yes.
******
Q. Assuming that the Census Bureau was unable to complete matching studies to determine the Hispanic undercount, but did have the black undercount through demographic analysis, in your opinion would it be reasonable to assume that the Hispanic undercount was substantially the same as the black un-dercount?
A. Well, if you had nothing else to go by, this is one judgment that could bring the statistics closer to the truth than the unadjusted data. In my judgment, yes.
Q. Would it also be reasonable, in light of what you have said, to assume that the His-panic undercount was intermediate between white and black, but closer to black?
*638A. That is another judgment, as to between these judgments, nobody would have a definitive answer as to which was the better judgment.
Q. In your opinion, from a demographic standpoint, would either judgment be a reasonable one?
A. Well, in light of what we know about the characteristics of the predominant proportion of the population of Hispanic origin, I would say either judgment would have some justification and some reason, and either one could conceivably bring the number closer to the truth than the unadjusted data. App. at 418-20.
Dr. Hauser testified that in his expert opinion the recent coverage improvement initiatives of the Census Bureau would create no net improvement in the differential undercount rate:
Q. By factors that would tend to increase the undercount, are you referring to the coverage improvement plans of the Bureau?
A. Well, factors — this would tend to decrease the undercount.
Q. Decrease the undercount?
A. Decrease the undercount. The Bureau has spent more money and procedures to try to minimize the undercount at this time than in any previous census. These would tend, if successful, would decrease the size of the undercount.
Q. I believe you indicated earlier that because of demographic considerations, there are, in other words, approximately 2,000,000 people likely to be undercounted if there is no improvement in coverage?
A. About an extra million or more by reason of just the increased proportions of blacks and increased proportion of hard-to-find youth, who are there because of the post-war baby boom.
Q. So that in lieu of that, is it possible that the coverage improvement features will be offset by this and so will not have any net effect?
A. Well, this and other factors of the type to which I referred. Remember this, a third of all blacks live in poverty by the government standards. Twenty-seven per cent of Hispanics live in poverty by government standards. The combination of their socioeconomic status, residences which are often of substandard quality, hard to locate, residential patterns of occupancy which make it very difficult to find separate households within cramped quarters, converted structures, some legally converted, some illegally converted, that the socioeconomic station of our minority groups is such that by reason of that status and other research by the Bureau tends to support this, that they are difficult to count. App. 403-5.
. See, e. g., the testimony of Dr. Robert Hill, App. 507-10; the testimony of Dr. Taeuber, App. 631-2; and the testimony of Dr. Hauser, App. 424-5, 444-5.
. For example, Dr. Hill testified in part as follows:
Q. Sir, do you have an opinion as to whether or not an adjustment of the 1980 Decennial Census to include the undercount would permit population totals based on the techniques that are available to the mathematical and statistical and demographic communities to arrive at population, corrected population totals, that are closer to the truth than the numbers that would be obtained from the counts?
A. I do have an opinion.
Q. What is that opinion, sir?
A. I do believe that the population estimate as derived by the demographic procedure at the national level will yield more accurate count of the population than will be derived by the enumeration procedures of the Census Bureau.
Q. Do you have an opinion as to whether there are techniques to distribute that under-count from the national level into the local areas so that the local areas which have population counts that are corrected would have counts that would be closer to the truth when comparing to those population counts that were not corrected?
A. Yes, I do have an opinion.
Q. What is that, sir?
A. I believe that, first of all, in terms of the state of the art at this time, that the demographic analysis yields the most accurate measure of the undercount at the national level. By utilizing the synthetic method for deriving estimates of the undercount for states, I think it is an extremely reliable procedure for deriving estimates of the under-count for states, so that by taking what I did from the Census Bureau demographic estimates at the national level, one can then derive reliable estimates of the undercount for states.
Q. How about below the state level?
A. Well,- now here I think that there are several options that are available. One is that I also believe that one can derive reliable estimates below the state level, but clearly demographic analysis tells us several things. One is that there is an inverse relationship *639with reliability and size of place, so that we would expect that the smaller the area, we will have greater error in the estimates that we are deriving. But I think for many large metropolitan areas you could use the synthetic method for being able to derive those. So 1 think that it can be used. There is another procedure which is what the Census Bureau has been using, which I think is a viable option, and it was particularly used in their [sic] where you counted procedures, where what you have in effect, if you take the state totals that are derived by the synthetic method, you can therefore, according to the procedures that they use and when you count it, you have a number of—
Q. (Interposing) Where you counted?
A. Where you counted, the similar procedure, the imputation procedures they used for those.
Q. In 1970?
Q. Do you have an opinion, if they used a synthetic method, taking the national under-count rates by race and applied the synthetic method from — down to the national level and to local areas to the degree that they have to report to the state on April 1st of 1981 for Congressional apportionment purposes that those numbers, just using a synthetic method alone, whether or not the corrected totals for the undercount would be on the average closer to the truth than the population totals that were not corrected?
A. I do have an opinion.
Q. And what is that?
A. It is that those estimates would be closer to the truth than unadjusted population counts for areas on the average.
Q. Would that be true for a significant proportion of the areas that were adjusted?
A. Yes, it would be true for the significant proportion of the areas that were adjusted.
Q. Would there be areas where there is a possibility that an adjustment for under-counts in some percentage slightly less accurate than the unadjusted population counts?
A. I would think that there would be, but I think that there would be a very small minority.
Q. Now with regard to using the synthetic method just for distributing population down to the local levels for apportionment purposes, with regard to the relative proportions, do you have an opinion as to the degree of reliability that would grow out of that type of an undercount adjustment?
A. Yes, I do have an opinion.
Q. What is that?
A. I think the real virtue, frankly, of the synthetic method is that in addition to giving you reliable estimates in an absolute sense, it is even much more powerful in giving you reliable relative proportions of the count relative — places in relation to each other which is the basis for apportionment.
Q. Explain to the Court what you mean by relative apportionment?
A. What I mean is the synthetic method would give you some estimates for areas. Now the real significance is not in some cases what those may be in their absolute numbers, although that is part of what you would derive, but even more significant is that you would consistently get estimates that were in the proper ordinal relationship to each other, that those places that in many cases we suspect would tend to have higher undercount rates because of these factors that we are talking about, would, by repeated different measures, would yield estimates higher than those that one would tend to have less of those characteristics, so that you would be able to distribute your entire population for a state in very reliable terms for any boundary distribution or what have you using the synthetic method.
. The original complaint in this case was filed on Census Day, April 1, 1980. Testimony was heard during August of 1980, and the district court opinion and judgment for the plaintiffs was issued on September 25, 1980. The release of significant amounts of data concerning the actual 1980 population count was not available until later. However, by October 27, 1980, preliminary field count figures were available for areas where 95 percent of the U. S. population resides. See Report Of The United States Bureau Of The Census To The Judgment Of The Court 2 reprinted in Appendix at 122.
. The relevant stipulation reads as follows:
The parties hereby stipulate the following facts are true:
A. In the taking of the 1980 decennial census, there will be a census undercount, in that all of the persons actually residing in the United States on the census date, April 1, 1980, will not be enumerated as part of the 1980 decennial census.
B. In the taking of the 1980 decennial census, it is likely that there will be a differential of completeness for blacks and Hispan-ics when compared to whites.
Final Pre-trial Order, para. 3, at 3, reprinted in Appendix at 61.
. At the time the Bureau filed its Post-Trial Report, its preliminary field count for the nation as a whole showed 225,590,000. By then, it was “increasingly confident that the final [head]count (would) . .. exceed 225 million and approach 226 million.” Pre-censual estimates of the actual 1980 population projected that about 221,672,000 persons would be within the country at that time. Post-Trial Report, at Appendix 129-33. This latter figure contained no “allowance for undernumeration in the previous Census.” Id. at 130.
. I am surprised by how apparently willing the majority is to accept the Bureau’s ex parte post-trial analysis. This stands in sharp contrast to the majority’s silence regarding the district court’s findings of fact after trial. The district court’s findings are supported by substantial evidence and are not clearly erroneous. In fact, the government has never challenged the district court’s findings on appeal. The government’s main brief argued only that 1) the plaintiffs lacked standing, 2) the Bureau’s decision to base the census on an actual enumeration was rational and not arbitrary, and 3) the district court abused its discretion in ordering injunctive relief. In its reply brief, the government, while arguing that no statistically defensible adjustment method exists, takes a curious position. It states:
In addition, at a number of places in their brief, plaintiffs claim that we cannot or do not dispute the findings made by the district court. As noted in our introduction, this claim is false. As we show below, the district court never should have reached these factual conclusions in the first place. Consequently, we attack them by demonstrating that the record proves that the Census Bureau conducted the census in a rational manner. Therefore, it is unnecessary for the Government to argue that the findings were clearly erroneous. Gov’t Reply Brief at 14-15.
As I note, I do not think that the district court applied an incorrect legal standard. Even if it did, however, the underlying factual findings are still subject to the clearly erroneous rule. Accordingly, the government’s attacks on the feasibility of adjustment must be tested under this rule. It is disingenious for the government to snipe at the district court’s findings while claiming that it does not have to challenge them as clearly erroneous.
. Evidence introduced in the Carey v. Klutz-nick trial apparently contradicts the Bureau’s conclusion that there are no available methods to measure the 1980 undercount. The trial court found:
While the “measured undercount” for the nation has been reduced to almost zero by the latest estimates, the Census Bureau has no evidence that there has been a decrease in the differential undercoverage of the groups that traditionally are “hard to enumerate.” There is, however, evidence from independent surveys that the populations of New *641York State and New York City which contain relatively large numbers of the “hard to enumerate” were undercounted in 1980 in differentially high proportions to that of the nation as a whole.
Conservative estimates of the undercount in 1980 for New York City ranged from 517,-000 to 650,000 and for New York State ranged from 772,000 to 905,000. Plaintiffs Ex. 235 at If 19. The results of a random phone survey of New York City and New York State residents conducted between August 20, 1980 and September 11, 1980 by the independent public affairs research organization, Penn & Schoen, indicate that 6% of the population of New York State and 8% of the population of New York City have not been counted by the Census Bureau. While the defendants have challenged the reliability of this survey I find that it is entitled to some weight in demonstrating the magnitude of the undercount in New York.
Other evidence of the undercount in New York are the results of “Were You Counted Campaigns” in New York City and Rochester. The results of the New York City “Were You Counted” campaigns show that 18,109 or 11.2% of the 161,718 applicants or recipients interviewed by 41 Income Maintenance Centers, 8 recertification offices and 16 nonpublic assistance Food Stamp Offices of the New York City Human Resources Administration had not been counted. A mail survey conducted of all 302,000 households in the New York City Income Maintenance Program indicated that of the 281,130 households that responded, 217,325 stated that they had returned a census form or were visited by a census taker, 41,175 did not answer the question and 22,270 or 8% of the respondents stated that they did not return a census form or were not visited by a census enumerator.
Defendant’s have submitted a preliminary report on the processing of “Were You Counted” forms for New York City. The results indicate that 9,796 or approximately two-thirds of the 15,158 individuals who claimed they were not counted lived in households from which the Census Bureau had already received a mail questionnaire or a field return questionnaire showing an equal or greater number of persons. Of the remaining one-third or 5,362 persons, 83.04 percent of the individuals contacted by the Census Bureau were found to have been missed by the census. Of 2,123 persons contacted by the Bureau 1,763 were added to existing census figures. If the 3,239 individuals who were not contacted by the Census Bureau are deemed to have been missed at the same rate as persons already reached by the Bureau, this would result in the addition of 2,688 (83% of 3,239) individuals to existing census counts. The total figures would then reveal that of 15,158 individuals who claimed they had not been counted, 4,451 or 29.4% were in fact not counted. Of course, many more individuals have submitted “Were You Counted” forms in New York City and approximately 30 percent of that population were most likely missed by the Bureau.
In the City of Rochester “Were You Counted Campaign,” 7 to 10 percent of those contacted stated that they had not been enumerated. Of those who claimed they had been missed by the Census Bureau, approximately 60% were verified by the Census Bureau as not having been enumerated. 508 F.Supp. 420, 428.