McNeil v. McDonough (515 F. Supp. 113, 1980)

OPINION

BIUNNO, District Judge.

This is a case in which the claims made involve allegations of racial discrimination in employment and employment opportunities. Trial was before the court without a jury. As provided by F.R.Civ.P. 52, this opinion contains the findings of fact and conclusions of law contemplated by the Rule.

The claims are asserted on a number of federal grounds with corresponding authority for jurisdiction, as follows:

Under Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e, jurisdiction being under 28 U.S.C. § 1343 or 42 U.S.C. § 2000e-5(f)(3), or both;

Under 42 U.S.C. §§ 1982, 1983 and 1988, with jurisdiction under 28 U.S.C. § 1343;

Under Amendment 14 of the U.S. Constitution with jurisdiction under 28 U.S.C. § 1331(a).

Different prerequisites, time limits or statutes of limitation apply to the different grounds, and different elements and defenses apply, as discussed hereafter.

The individual defendants originally named were formerly in office as Director of the Division of Alcoholic Beverage Control in the Department of Law and Public Safety of the State of New Jersey. The present Director in office, Joseph H. Lerner, was impleaded by agreement of the parties as though originally named as a defendant and as though his answer were the same as the others. The court notes that the method provided for by F.R.Civ.P. 25(d)(2), i. e., describing the public officer by title rather than by name, was not used. The final defendant is the Division of Alcoholic Beverage Control itself.

The plaintiffs, Wilbur J. McNeil and William Jones work in the Division.

Before launching into an analysis of the law and testimony, it will be useful background to touch on the history of the Division itself.

The Division is the current administrative successor to the N.J. State Department of Alcoholic Beverage Control, established by N.J.P.L.1933, c. 436 as a consequence of the adoption of Amendment 21, which in turn repealed Amendment 14 (Prohibition) of the U.S. Constitution. The history of regulation and control of alcoholic beverages, however, goes back much farther than that, perhaps as far back as Allinson’s Laws. (1738-1739) Game v. Burnett, 122 N.J.L. 39, 4 A.2d 37 (S.Ct.1939) reviews this long history before federal Prohibition and Repeal.

The 1933 statute followed an exhaustive study and report by a committee, established by joint resolution of the Legislature adopted October 9, 1933, and largely followed the draft statute proposed by the report.

As was observed in an earlier study, “It was felt that the choice of a ‘single executive’ was a wise one. The object of liquor legislation is rigid, uncompromising control. The main duties of the department are police in nature, and such duties are performed with greater efficiency and promptness through a single executive than through a plural board.”

Study of State Administrative Agencies in New Jersey, by Clark Crane Vogel under the supervision of Nathan L. Jacobs (University of Newark, May, 1941), at p.44.

With the enactment of the Revised Statutes of 1937, the statute became N.J.R.S. Title 33. The original statute, by § 3 (N.J. R.S. 33:1-3) not only created the Department, but itself named the first commissioner, D. Frederick Burnett, Esq., and fixed his term and salary. Mr. Burnett had served as counsel to the study committee (see Study supra, p.43 at n.9). It is a matter of common knowledge among the State’s bench and bar that Mr. Burnett was not only the architect of the statute but that as its first commissioner, and with the able assistance of his Chief Deputy Commissioner and Counsel, Nathan L. Jacobs (until lately Associate Justice of the Supreme Court of New Jersey) set the level of high professional performance called for by the command that he do all things “designed to insure the fair, impartial, stringent and comprehensive administration” of the law. N.J.P.L. 1933, c. 436, § 20; N.J.R.S. 33:1-23.

The structure of the department as created did not contemplate that it should itself be the sole vehicle for control and administration. Rather, a large measure of “home rule” was delegated to local municipalities with the Department serving as supervisor. The commissioner was assigned the function of issuing certain types of licenses (e. g., for manufacture, wholesale, plenary retail transit, transportation and public warehouse), N.J.R.S. 33:1-18, while local municipal boards or bodies were authorized to issue all other kinds of license, mainly various types of retail licenses which are far more numerous, N.J.R.S. 33:1-19.

With this municipal licensing authority there was imposed a set of corresponding administrative enforcement duties, N.J.R.S. 33:1-24, including the suspension or revocation of municipally issued licenses, NJRS 33:1-31, subject to appeal to the commissioner. As of June 30,1938, a total of 3,943 licenses had been issued, Study supra, at p. 51, most of which doubtless were municipally issued retail licenses.

The structure, then, with the large measure of “home rule” for retail licenses, implied that the Department was intended to be a small, highly organized and efficient unit, directly controlling the manufacture, wholesaling and distribution aspects, while auditing the municipal functions through appeal and otherwise.

With the adoption of the 1947 N.J. Constitution, it was required that all executive branch units, including departments, be allocated by law among and within not more than 20 principal departments, N.J.Const., 1947, Art.5, § 4, par.1. Under current law the former Department has been constituted as a division of the Department of Law and Public Safety, headed by the Attorney General, N.J.R.S. 52:17B-15, and the powers and duties of the former commissioner transferred to the Director of the Division, N.J.R.S. 52:17B-17.

Given this background and pattern, it is no surprise that the number of budgeted positions in the division is only about 150, a small number for a unit of State government charged with the important functions and duties of this one. Some of the positions are clerical force, which come under Civil Service (NJRS Title 11), by reason of N.J.R.S. 33:1 — 4c. All other State employees appointed by the director do not come within Civil Service. These fall into a number of categories:

... 5 deputy directors, each assigned to a bureau by the director, and removable for cause; N.J.R.S. 33:1-4 b (1973); ... such number of inspectors, investigators and executive assistants as the director deems necessary; they are removable at will, except that after 3 years employment they shall serve during good behavior and shall not be removed except for cause; Idem., par. d. ... temporarily employed experts and specialists for specified service;

... counsel and other legal assistants.

Of the total staff, the evidence is that there are 80 to 90 employees in the Enforcement Bureau, these being divided into 5 sections: wholesale, retail, undercover, front and administrative. McNeil has served in both the undercover and retail sections and is now assigned to wholesale. Jones has been continuously assigned to undercover.

Although the original commissioner was named by the Legislature, it may no longer elect or appoint any executive, administrative or judicial officer except the State Auditor, N.J.Const., 1947, Art. 4, § 5 par. 5. The director of the Division is now appointed by the Governor, with the advice and consent of the Senate, N.J.R.S. 52:17B-16. The office was made full-time by N.J.P.L. 1970, c. 222, effective October 16, 1970, Idem, (pocket part).

Thus, the Director of the division is today a state officer appointed by the Governor and confirmed by the Senate, whose immediate head is the Attorney-General of New Jersey, who in turn, as head of the Department of Law and Public Safety, is a member of the Governor’s cabinet. In this capacity he serves an administrative state function separate from his duties and powers as chief law enforcement officer for the State as a constitutional officer, without regard to departmental structure or administrative duties. See N.J.Const. 1947, Art. 5, § 4, par. 1, 2 and 3; also State v. Winne, 12 N.J. 152, 96 A.2d 63 (1953); Board etc. Lehigh Valley R. Co., 106 N.J.L. 411, 149 A. 263 (E & A, 1929).

Limitations Aspects

There is no federal statute of limitations for suits under the Civil Rights Act, 42 U.S.C. § 1981 et seq. The decisions widely recognize that in this circumstance the federal courts are to apply that period of limitation applicable under State law had a like action been brought under State law. In fact, the conformity provisions of 42 U.S.C. § 1988 contemplate that this course be followed.

Inspection of the reported cases dealing with the selection of the proper State limitation period show wide variations of scope. As observed recently in Davis v. U. S. Steel Supply, 581 F.2d 335 (CA 3, 1978) the selection involves a process of analyzing each complaint, and each aspect of a particular complaint, to ascertain the fundamental nature of each claim since different time periods may apply to different claims or complaints.

In this case, both plaintiffs have been and are still employees of the State of New Jersey, albeit in its unclassified service rather than under Civil Service, and complain of invidious discrimination grounded on race or color in the alleged failure to promote each of them. Since each has 3 years’ service, neither may be discharged without cause under present local statute.

Accordingly, the claim may be read as asserting that the employment relation, which is contractual in nature, is to be regarded as though it contained a provision that the employer would not act contrary to applicable law. Such a reading analogizes the claim as one for breach of contract, to which New Jersey’s 6-year statute of limitations, N.J.S. 2A:14-1 would apply. That statute will be taken as governing all claims encompassed by 42 U.S.C. § 1981, et seq., as well as claims said to arise directly under the U.S. Constitution, in respect to which the same underlying considerations apply.

Plaintiff McNeil filed his complaint on May 28, 1976, and for claims under the Civil Rights Act and the 14th Amendment encompasses alleged conduct back to May 29, 1970 but not earlier.

Plaintiff William Jones was allowed to be added by the pretrial order dated December 16,1976 and so his corresponding claims will reach back to December 15, 1970 but not before.

The present Director, Joseph H. Lemer, was added as a defendant by amendment filed October 16, 1978. As to him, such claims of both parties reach back to October 15, 1972 but not before assuming he was in office then.

In the case of the claims under 42 U.S.C. 2000e (the Civil Rights Act of 1964) the basic claim arises from the definition of what is an “unlawful employment practice for an employer”, 42 U.S.C. § 2000e-2(a). The provisions for mechanisms to prevent unlawful employment practices are found in § 2000e-5. Charges are to be filed within 180 days of the alleged unlawful practice, with EEOC; but if the matter begins with a proceeding filed with a local agency of the same kind as EEOC, then the time for filing with EEOC is either 300 days after the alleged unlawful practice, or within 30 days after notice that the local proceeding has terminated, whichever is earlier; idem, par. (e). Assuming these time limitations with EEOC are satisfied, then the person aggrieved may file a civil action (either in a US District Court or a State court, see Peper v. Princeton, 77 N.J. 55, 389 A.2d 465 (1978); and if the Attorney General (US) has not filed a civil action in a case involving a government or governmental agency within 180 days of the filing of the charge with EEOC, the person aggrieved is to be so notified, and may bring the civil action within 90 days after the notice.

McNeil filed his Title VII charge with EEOC on May 25, 1974, and was issued a right to sue letter under date of March 4, 1976. Complaint was filed May 28, 1976, within 90 days, and can reach back 180 days before the EEOC filing, or to November 26, 1973.

Jones filed his EEOC complaint February 18, 1975 (P-36). His “right to sue” letter from the Department of Justice is dated October 1, 1976, and if received the same day, the 90 days would have expired December 31, 1976. He was joined as plaintiff on December 16, 1976 and so his claim under 42 U.S.C. § 2000e is not barred by limitations. His claim can reach back to 180 days before February 18, 1975 (EEOC filing date), or to August 22, 1974.

The amendment removing the exemption of state and local governments from Title VII had an effective date of March 24,1972, and alleged violations after that date may be asserted, subject to the applicable limitations. See the Equal Employment Opportunity Act of 1972, Pub.L.No. 92-261, 86 Stat. 103.

The Immunity Question

The complaint, as amended, names as defendants the present director of the ABC and three former directors (McDonough, Bowers and Ronco), individually and as Directors, plus the “State of New Jersey Division of Alcoholic Beverage Control.”

From the legislative history outlined above, there can be no doubt that insofar as the ABC is itself named as a defendant, the State of New Jersey is effectively named. The ABC is a division within the Department of Law and Public Safety, headed by the Attorney-General, and one of the principal departments in the Executive Branch. The employer is the State itself. Plaintiffs are state employees in the unclassified service. This fact being undeniably so in light of the structure of state government as reflected in its local Constitution and statutes, it follows that the only claims for money damages that plaintiffs can. assert against the State are those coming within 42 U.S.C. § 2000e, et seq., because the 1972 amendment contained a Congressional waiver of the State’s immunity from suit under Amend. 11, through the exercise of its legislative power to implement. Amend. 14 under § 5 thereof. Thus, no monetary claim can be asserted against the State under 42 U.S.C. § 1981, et seq., because for those statutes there has been no waiver of sovereign immunity, either by Congress or by.the State. See: Monroe v. Pape, 365 U.S. 167, 81 S.Ct. 473, 5 L.Ed.2d 492 (1961), as modified by Monell v. Department, etc., 436 U.S. 658, 98 S.Ct. 2018, 56 L.Ed.2d 611 (1978); Parden v. Terminal Railway Co., 377 U.S. 184, 84 S.Ct. 1207, 12 L.Ed.2d 233 (1964); Edelman v. Jordan, 415 U.S. 651, 94 S.Ct. 1347, 39 L.Ed.2d 662 (1974); Fitzpatrick v. Bitzer, 427 U.S. 445, 455, 96 S.Ct. 2666, 2671, 49 L.Ed.2d 614 (1976); Cannon v. University of Chicago, 441 U.S. 677, 99 S.Ct. 1946, 60 L.Ed.2d 560 (1979); Chapman v. Houston, etc., 441 U.S. 600, 99 S.Ct. 1905, 60 L.Ed.2d 508 (1979); and the analysis in Skehan v. Board, etc., 590 F.2d 470 (CA 3, 1978).

Of course, if a case is made out on the merits, the 11th Amendment does not stand in the way of prospective injunctive remedies to the extent they are otherwise warranted.

The Substantive Law

As the decisions in this evolving field of employment discrimination seem to indicate, the substantive law is essentially the same without regard to whether the claim is asserted under 42 U.S.C. § 1981, et seq., or under Title VII, 42 U.S.C. § 2000e, et seq.

In terms of “essential elements” that must be established, of course, these will vary with the applicable statute. Thus, under 42 U.S.C. § 1981, the sole issue is “race”, i. e., whether the claimant has been deprived of the same enumerated rights “as are enjoyed by white persons.” Under 42 U.S.C. § 1983, the issue is whether the claimant, as a citizen of the United States, has been subjected under color of local law to the deprivation of any rights, privileges or immunities secured by the federal Constitution and laws. Under 42 U.S.C. § 2000e, et seq., the denounced grounds of treatment which make an employment practice unlawful are those based on race, color, religion, sex, or national origin. See 42 U.S.C. § 2000e-2.

In the case of non-federal employment (which now embraces traditional “private sector” employment as well as employment by a non-federal government, government agency, etc.) a further distinction is made by 42 U.S.C. § 2000e-5(f)(l) in respect to prerequisites to suit. For strictly private sector claims, it is EEOC which can file suit or issue a 90-day right to sue letter. For state government and state agency claims, that function is assigned to the Attorney General (Department of Justice).

Regardless of what the forbidden criteria are under a given statute, it seems well recognized that four elements must be shown to make out a prima facie case of discrimination:

1. the claimant falls into one of the forbidden categories, as may be applicable under the particular statute involved;

2. the claimant applied and was qualified for a job (including a promotion) which was vacant and which the employer was seeking to fill;

3. that although the claimant was qualified, he was rejected;

4. that after rejection of claimant, the employer continued to seek applicants who possessed claimant’s qualifications.

See, McDonnell Douglas v. Green, 411 U.S. 792, at 802, 93 S.Ct. 1817, 1824, 36 L.Ed.2d 668 (1973).

Thus, a case based on race discrimination might arise under § 1981, § 1983 or § 2000e, or all three. A case based on freedom of religion may arise under § 1983 even though the employment relation is not involved, and may arise under both § 1983 and § 2000e if it is involved. But a case grounded on gender discrimination (at this stage of the law) can only arise in the employment context and only under § 2000e.

These examples aside, however, the four elements of a prima facie case under McDonnell Douglas probably are not a closed class or enumeration since the ultimate issue in every case of this kind is whatever conduct the applicable statute may denounce. The four elements of McDonnell Douglas appear to be no more than an articulation of the kind of showing that is to be accepted by the courts as making out a prima facie case, i. e., such that if the evidence is closed at that point, the claim must go to the finder of fact (jury or judge) who may, but is not obliged to, find for the claimant on the showing made. The level of proof called for, in light of the ultimate issue to be decided, is clearly intended to be somewhat lower than would be called for in some other kind of case, in recognition of the difficulties of adducing more direct forms of evidence. It might well be said that the four elements of McDonnell Douglas represent a set of circumstances regarded by law as sufficient to make a prima facie case.

They cannot be regarded as an exclusive or closed set of elements, however, because (1) each case will necessarily turn on its own facts; (2) as articulated they do not reach the discriminatory discharge type of case; (3) they are probably inextricably intertwined with the law on order and allocation of proof rules, which have been developed to deal with all cases in the general field.

Kinds of evidence to be considered

The development of a body of law through the decisional process is one based on reason and experience. It begins by reviewing and analyzing the set of existing decisions, and from that point developing an expression of a general principle by the logical process of induction. The general principle so developed is then applied to the decision of new disputes, each having its own set of facts, by the logical process of deduction.

While it may be oversimplified, the history of modern jurisprudence shows that this process of induction and then deduction represents an attempt to apply to that part of the social sciences with which the law and the judicial process are concerned, the methods which have been so successful in the natural sciences from the days of the Renaissance and before.

The behavior of water when heated, is a good example. Many experiments by different investigators, for example, showed a direct relation between the quantity of heat applied and the extent to which the temperature of the water rose. These and other observations led to the development of the theory of heat, before which it was only known that an object was (relatively) hot or cold. The study of observations led to the establishment of identifiable and precise units for the measurement of quantities of heat: the calorie is defined as that quantity of heat that will raise the temperature of one gram of water by one degree Centigrade, while the British Thermal Unit (B.T.U.) is that quantity of heat that will raise the temperature of one pound of water by one degree Fahrenheit.

From like observations, the temperature at which water boils was found to be a constant, so long as the water is pure, depending on the pressure at the surface. At normal pressures and at sea level, the temperature is 100° C or 212° F.

It was found that to melt a given quantity of ice, a change of state that does not raise the temperature, 80 times as much heat is needed as would be required to raise water temperature by one degree. This is the “heat of fusion” for ice. At the other end of the scale, where water is converted into steam, it was found that 537 times as much heat was needed to convert a given quantity of water into steam without raising its temperature or pressure. This is the “heat of vaporization” of water.

These physical “laws”, derived by induction from many experimental observations, and others like them, have led to countless discoveries and formulations in the natural sciences by applying the general principles so induced to specific other states of fact by deduction.

Although the legal decisional process is analogous to that of the scientific method, and deliberately so designed, it is not the same. Water will boil at whatever temperature it does according to natural forces. The most that can be done is to describe it. The forces cannot be altered. In the social sciences, this is not true or, if it is true no means have been developed to achieve the same level of confidence and predictability. The law, as a branch of the social sciences, necessarily deals with the conduct and behavior of human beings, both as discrete individuals and as interacting populations of discrete individuals. About all that is known is that the behavior of discrete individuals or of populations cannot be forecast reliably in the same sense that the behavior of substances and forces can be in the natural sciences.

These observations are considered to be of considerable importance in an evolving field of law such as this because, to quote Mr. Justice Holmes, “The life of the law is not logic, it is experience”, (emphasis added). New decisions, like new scientific experiments, are the raw material by which the law, or a scientific principle, constantly evolves, changes or is refined. He did not discard logic; he emphasized the importance of experience. Logic is merely the tool by which experience is evaluated, understood and applied.

Keeping in mind the cautions just discussed, the reported cases in the field so far recognize five categories of basic circumstances that will make out a discrimination in employment case:

1. individual disparate treatment cases, where the claimant adduces evidence which, if believed, tends to show less favorable treatment than peers due to some applicable forbidden criterion such as sex, race, etc. The prime example here is McDonnell Douglas v. Green, 411 U.S. 792, 93 S.Ct. 1817, 36 L.Ed.2d 668 (1973);

2. pattern and practice cases, where the claim is one of systematic disparate treatment of a category to which the claimant belongs, even though he cannot show “individual disparate treatment” directed to him. The major example of this category is Hazelwood School District v. U. S., 433 U.S. 299, 97 S.Ct. 2736, 53 L.Ed.2d 768 (1977);

3. disparate impact or effect cases, where the claim is that some seemingly neutral test for employment (hiring or promotion) criterion disproportionately disqualifies individuals in a category protected by the forbidden criteria, and that the test or criterion is not justified by business necessity. The lead case for this category is Griggs v. Duke Power Co. 401 U.S. 424, 91 S.Ct. 849, 28 L.Ed.2d 158 (1971). A later case in this group is Dothard v. Rawlinson, 433 U.S. 321, 97 S.Ct. 2720, 53 L.Ed.2d 786 (1977);

4. perpetuation of discrimination cases, where the claim is that the application of a seniority system carries into the future the effect and impact of discriminatory practices preceding the effective date (in the particular case) of Title VII. These mainly arise out of exclusion practices of labor unions (though not exclusively). One example out of many where this element was involved is U. S. v. Int’l Union of Elevator Constructors, 538 F.2d 1012 (CA 3,1976). However, the decision in Int’l Brotherhood of Teamsters v. U. S., 431 U.S. 324, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977) seems to narrow this category to those cases where the seniority system is itself discriminatory or had its genesis in a now-forbidden kind of discrimination (e. g. racial).

5. Affirmative action cases, which may or may not involve prior discrimination of any kind, where a program is adopted to try to overcome current or expected disparities. Examples here are University, etc., v. Bakke, 438 U.S. 265, 98 S.Ct. 2733, 57 L.Ed.2d 750 (1978) and United Steelworkers v. Weber, 443 U.S. 193, 99 S.Ct. 2721, 61 L.Ed.2d 480 (1979).

These five categories are no more than an attempt at a rational grouping of types made up of the very large number of reported decisions at all levels. Before Bakke and Steelworkers there were only four classifications. Next month there may be a sixth. The classification is not closed, nor can it be closed any more than classifications of “fraud”. Courts have always refused to express any closed classification of all types of frauds, because if they did, an ingenious schemer would find some other way to defraud that was outside the definition of classifications or categories. The same is obviously true of cases involving claims of forbidden discrimination in employment.

There are also cases arising under other statutes, and not involving “civil rights” in the sense discussed. A good example is U. S. v. Testan, 424 U.S. 392, 96 S.Ct. 948, 47 L.Ed.2d 114 (1976) where trial attorneys in a federal unit that involved matters coming before the Armed Services Board of Contract Appeals of the Department of Defense were classified by the U.S. Civil Service Commission in grade GS-13, and complained that their duties were identical to those of attorneys in the Contract Appeals Division (Judge Advocate — Air Force) who were classified in grade GS-14 (a higher compensated level). The decision turns on collateral matters such as the jurisdiction of the Court of Claims, and the merits of the claim are not decided. Yet, though no issue of race, color, religion, sex, national origin, etc., is mentioned, an employee in GS-13 who feels he belongs in GS-14 has exactly the same reaction and view of being “discriminated against” as a claimant under § 1981, § 1983 or § 2000e.

Order and Allocation of Proof

This aspect of cases involving claims of unlawful discrimination in employment is unique only because the Supreme Court of the United States has found it necessary to indicate the kinds and quantum of evidence needed to adduce a prima facie case in this developing field of litigation. The concept that a claimant advancing an affirmative claim carries the burden of persuasion for the whole case is hardly novel. The notion that he must adduce some specified minimum, more than a modicum or scintilla, to survive a motion for involuntary dismissal at the end of plaintiff’s case in a bench trial [Rule 41(b), F.R.Civ.P.] or a motion for a directed verdict in a jury trial at the same stage [Rule 50(a), F.R.Civ.P.] is also well established.

Similarly, a defendant who rests on the motion (either kind) without tendering evidence of his own to meet a prima facie case takes the risk that the finder of fact, as distinguished from the judge ruling on a question of law, may conclude that the prima facie case also carries the burden of persuasion. It may be that the finder of fact will not be so persuaded, but a defendant who offers no evidence to meet what the judge concludes is a prima facie case as a matter of law takes the risk of the ultimate finding.

The line of cases on this subject runs from what was said in McDonnell Douglas v. Green, 411 U.S. 792, 93 S.Ct. 1817, 36 L.Ed.2d 668 (1973), through Furnco etc. v. Waters, 438 U.S. 567, 98 S.Ct. 2943, 57 L.Ed.2d 957 (1978) to Board etc. v. Sweeney, 439 U.S. 24, 99 S.Ct. 295, 58 L.Ed.2d 216 (1978), and the decisions are by divided votes.

What they say, in substance, is that which is generally understood, namely that once a claimant makes out a prima facie case (in this category under specifically articulated standards), the burden of producing evidence (but not the overall burden of persuasion) shifts to the defendant, who must come forward with the articulation, by evidence, of some legitimate, non-discriminatory reason for the conduct complained of, and need not do more, though he may if he chooses.

The trial court is then obliged to consider and evaluate the evidence, including matters of credibility, to decide: first, whether the claimant has made a prima facie case under applicable standards; second, whether the defendant has adduced evidence to articulate a legitimate, non-discriminatory reason by way of explanation; third, on all the evidence as a whole, has the claimant carried the overall and ultimate burden of persuasion.

It does not matter what the order or sequence of evidence is. The three steps by way of evaluation are to be separately considered and decided at the conclusion of all the evidence.

On its face, this three-step process of marshalling the evidence and evaluating it is no different than what is done in every bench trial, or should be done in every bench trial by way of compliance with the spirit and purpose of Rule 52(a), F.R.Civ.P. The approach provides the Court of Appeals with both an evidential record and a careful articulation of findings of fact and conclusions of law such as to enable it to address directly and finally decide whatever issues are raised on appeal, and tb enable it to apply soundly the rule that a trial court’s findings of fact will not be disturbed unless clearly erroneous.

As is evident from what has been said, the three stage evaluation is significant and must be carried out only in cases tried to the judge without a jury. In a case tried to a jury, the verdict would stand, whatever it might be, so long as the jury were properly instructed on the law, which hardly would include such notions as a prima facie case, or the shifting burden to produce evidence. The jury would be instructed that the claimant carries the burden of persuasion by a preponderance, and that the defendant carries the burden of persuasion by a preponderance for any affirmative defenses. It should be noted that both Fumco and Sweeney were cases tried to the judge without a jury, and they accordingly apply here. The guidelines for instructing a jury in a case of this kind have not been articulated by the Supreme Court in any of the cases in the line.

The cases also imply that in cases of this type that are tried to the court without a jury, the trial court as a finder of fact is expected to be sensitive to the difficulties of making a case by any substantial kind of direct evidence. It should be not unwilling to draw inferences from circumstantial evidence, particularly where different circumstances tend to confirm each other and point to a particular outcome. No doubt there are discrimination cases where direct evidence is available; the court has heard some of these, and the experience is that they do not go to trial but are settled.

Aside from those, the task of sifting the evidence to distinguish a valid claim of discrimination from a groundless one can be a difficult task, especially where the ultimate issue of unlawful discrimination is masked by ingeniously subtle devices. Yet, the task must be performed without crossing the undefined boundary between sensitivity for a claimant’s 'difficult proof problems and outright bias or prejudice in favor of the claimant.

The problem is one that is typical of the difference between jury and non-jury trials. In principle, the review of a final judgment ought to be the same in either case. The fact and experience is that appellate courts, both federal and state, display a greater readiness to intrude into the fact-finding function in non-jury cases than they do in jury cases. This is especially true for those appellate courts, like New Jersey’s, which are expressly vested with original jurisdiction “as may be necessary to the complete determination of any cause on review”, N.J. Const., 1947, Art. 6, § 5, par. 3. Similar authority is not expressly provided for in the federal scheme, where the test in a non-jury case is whether a trial judge’s findings of fact must be left undisturbed unless “clearly erroneous”. See Sweeney v. Board, etc., 604 F.2d 106, at 109 n.2 (CA 1, 1979); Scott v. Univ. of Delaware, 601 F.2d 76, at 81 (CA 3, 1978), cert. den. 444 U.S. 931, 100 S.Ct.2d 275, 62 L.Ed.2d 189 (1979); Rochez Bros. v. Rhoades, 527 F.2d 880, at 887 (CA 3, 1975); Kunda v. Muhlenberg, etc., 621 F.2d 532, 534 (CA 3, 1980). And compare Worthy v. U. S. Steel, 616 F.2d 698 (CA 3, 1980).

The point is not idly discussed, because there has been serious periodic debate over the question whether American courts should follow the English experience, where all civil cases are tried to the bench without a jury except for those involving fraud or defamation (see, e. g., “Should The Right to a Trial By Jury be Abolished in Most Types of Civil Cases?” 83 NJLJ 465; “Do We Need Juries In Civil Cases?”, 83 NJLJ 513; “Trial By Jury in Civil Cases”, 83 NJU 513; “On Juries in Civil Cases”, 87 NJLJ 268; “The Jury and Civil Actions in England”, 87 NJLJ 753.

One of the points made is that if all or most civil cases were tried without a jury, appellate courts would be more free to develop and lay down more specific and precise guides for trial judges than is fair in a mixed system. As it is, where some cases are tried to a jury and others not, depending on the choice of the parties, principle would reject the notion of sophisticated rules for bench trials only. The review function should be the same for both when both modes of trial exist.

Yet, even for classes of cases for which there is no right to jury trial (e. g., equity cases, trials of title in condemnation cases, and trials under the Federal Tort Claims Act), no studies of this aspect appear to have been made and no distinct body of law appears to have been accumulated from which one or more general principles can be derived by induction so that they can then be applied by deduction to new disputes.

In sum, as this court understands its function under the decisions referred to, it is to undertake the task of a sophisticated jury by evaluating the evidence through the three stages mentioned, and returning what would amount to a series of special verdicts.

This “jury-non-jury” analysis has been undertaken to aid in evaluating the cases which deal with concepts such as a “primafaeie” case, shifting of the burden to offer evidence, the risk of non-persuasion and the overall burden of persuasion because they have dealt with these concepts in non-jury cases like this one, although the concepts themselves have an underlying meaning or function which is significant only when there is a separation as between judge and jury, one dealing with law and the other with facts.

It is observed in passing, too, that the Supreme Court has flatly said that there is no right to jury trial in cases under Title VII, 42 U.S.C. § 2000e, Great American etc. v. Novotny, 442 U.S. 366, 99 S.Ct. 2345, 60 L.Ed.2d 957 (1979). This proposition was set forth with the rationale that “the Act expressly authorizes only equitable remedies”. This can only refer to 42 U.S.C. § 2000e-5(g), which mentions such compensatory monetary items as “back pay” as forms of “equitable relief”. See also, Richerson v. Jones, 551 F.2d 918, at 926 et seq. (CA 3, 1977), a case involving claims by a federal employee under 42 U.S.C. § 2000e-16, where punitive damages were ruled not allowable under Title VII because such damages are traditional in courts of law but not in courts of equity, relying on 42 U.S.C. § 2000e-5(g). Richerson is cited in Great American at footnote 17 for the same proposition.

Yet, under Rule 18, F.R.Civ.P., a party may assert as many claims, “legal, equitable, or maritime, as he has against an opposing party.” (Emphasis added) New Jersey has gone well beyond permitting joinder; it requires joinder of claims under the “entire controversy” doctrine now formally expressed in N.J.Court Rule R. 4:27-1(b) (1979), and see Pressler annotations to the Gann Law Books edition, 1980, for the major decisions on the subject.

Thus, a party may assert not only a claim for equitable (and hence non-jury) relief under Title VII, but also a claim under 42 U.S.C. § 1981,1983, etc., for which he may demand a jury. Such claims are asserted in this case, but since no jury was demanded by any party, the jury aspect was waived, Rule 38, F.R.Civ.P. and General Rule 8-F of this District. This is not a case like Great American etc. v. Novotny, supra, a case that went up from this circuit, where the Court of Appeals held that a claim under 42 U.S.C. § 2000e-3(a) [sec. 704(a) of Title VII] had been asserted — a non-jury claim not before the Supreme Court — but where the Supreme Court reversed the Circuit and ruled that a claim under 42 U.S.C. § 2000e-2(a) could not be asserted through an alleged conspiracy under 42 U.S.C. § 1985(c), the latter being a claim for which a jury trial might have been demanded.

For cases tried to a jury there are further implications. To illustrate: if a trial judge erroneously denies a motion for directed verdict under Rule 50(a) at the end of plaintiff's evidence, and decides that a prima facie case has been made out as a matter of law, all that the erroneous ruling implies is that on the evidential record at that stage there is a jury question. If defendant then chooses to rest and offers no evidence, the jury receives the case to decide on the facts. The judge’s erroneous ruling, in this example, is supposed to be made by accepting the plaintiff’s evidence as true, together with all favorable inferences that can reasonably be drawn therefrom by a reasonable jury. Even assuming the judge’s ruling to be wrong, the jury is in no way bound to accept all the plaintiff’s evidence, nor is it required to draw all the favorable reasonable inferences. On a plain a priori basis, the jury can return a verdict for the plaintiff or for the defendant. If the verdict be for plaintiff on the assumptions made here, the erroneous ruling can be corrected either by a motion n.o.v. under Rule 50(b), or on appeal from the final judgment.

But if, on the other hand, the jury verdict is for defendant, the judge’s erroneous ruling against defendant is mooted by the verdict and ensuing judgment. The error has been “cured” by the jury’s verdict leaving nothing to be appealed in respect to this point. It is for this reason that the subject is not one found discussed in the opinions of upper courts. The point can only arise on appeal if the verdict is for plaintiff in such a case and if the error is not corrected by judgment n.o.v.

The same observations apply in jury cases when defendant does offer evidence, including cases where plaintiff adduces rebuttal proofs, and the motion for directed verdict comes at the end of the evidence adduced by both sides. The only difference is that if there were a deficiency in plaintiff’s case in chief, such that it failed in law to amount to a prima facie case, the deficiency may be supplied through evidence of defendant’s testimony, whether on direct or cross-examination. Such an event, of itself, would “wash out” or cure the hypothetical erroneous ruling on the law at the end of plaintiff’s case.

In criminal cases, under local New Jersey law, this feature is different. A motion for acquittal at the end of the State’s case, if denied, is judged on appeal only according to the evidence adduced up to that point, and nothing appearing as evidence proffered by defendant may be considered on the point. The rule in federal criminal cases has been different, and the “jury question” issue is tested by all the evidence on both sides at the end of the entire case, although some indications suggest a change in the direction of the New Jersey rule may be developing.

In a routine, run-of-the-mill civil case tried to the court alone, whether it be by waiver of jury trial, or because it be one for which no right to jury trial exists, the court is called upon to make findings of fact and conclusions of law, Rule 52(a), and the same rule provides that:

“If an opinion or memorandum of decision is filed, it will be sufficient if the findings of fact and conclusions of law appear therein.”

Both requirements are, of necessity, matters of adjective and procedural law. This is for the obvious reason that the authority for the promulgation of Rule 52(a), along with all the other Rules of Civil Procedure, is derived from 28 U.S.C. § 2072, the second paragraph of which declares that:

“Such rules shall not abridge, enlarge or modify any substantive right and shall preserve the right of trial by jury as at common law and as declared by the Seventh Amendment to the Constitution”.

It is not to be assumed, in construing the Rules, that the Supreme Court has exceeded the scope of its authority. Rather, the Rules are to be interpreted in such fashion as will fall within that scope.

These well established principles and practices, however, have not been clearly applied in employment discrimination cases, whether under Title VII or § 1981, et seq., even in non-jury cases. Because of the uncertainty in the law and the absence of reliable guides to trial judges in these matters, the evaluation of the evidence will be made both as a matter of law (i. e., as though there had been a jury, and as though the court were ruling only on the legal sufficiency of the evidence), and separately as a finder of fact.

Kinds of evidence in discrimination cases

Despite the many analyses of the wide variety of discrimination cases, the outcome of all of them inevitably depends on the element of intent, and a causal relation between that intent and the consequence complained of. The “intent” referred to is not by any means confined to actual, subjective, individual intent of a purposeful nature. It refers to the legal concept of intent which also embraces subconscious and institutional intent as well as an intent inferred from purposeful adherence to some direction or course of action even after an invidious discriminatory effect has been shown to follow from what may have been wholly innocent conduct. Equitable concepts carry considerable weight here, for the ancient rule has been that where a course of conduct, innocently undertaken, is shown to invade the rights of others and the fact is brought to the actor’s attention, a refusal to alter the course of conduct to accommodate the rights of others will be regarded as deliberate and intentional. The principle is much the same as the even older concept, developed in the courts of law, that an entry by lawful authority will be treated as a trespass ab initio when the authority is abused after entry. The doctrine is said to have been first formally expounded in the Six Carpenters’ Case, 8 Coke 148. It was said in that case that the reason for the rule was that the subsequent illegal conduct shows that there was an “unlawful intent” to begin with.

The legal concept of “intent” is to be distinguished from “motive”, as juries in criminal cases are regularly instructed. Forbidden discrimination in employment may be the consequence of the highest and most salutary motives, but the legal “intent”, not the motive, is what controls.

Categorization of various kinds of evidence which the decided cases show may be accepted as proof of invidious discrimination must be understood as not implying that no other kinds of evidence may be used, or that several kinds of evidence may not be combined to prove the claim even though each kind, taken alone in a given case, may be insufficient as a matter of law or as a matter of fact.

It cannot be emphasized too strongly that each case will necessarily be evaluated on its own facts, and that no two cases will be exactly alike. Of course, there will be classes of cases, such as those where a large number of employers apply practices specified by an industry-wide collective bargaining agreement, in which the outcome of one dispute may have the practical effect, or may amount to binding precedent under the doctrine of stare decisis, of controlling the outcome of a very large number of other potential claims that never reach any administrative agency or court. The outcome of one case can oblige all the affected employers and collective bargaining agents to so modify the industry agreement as to eliminate any basis for other claims. This is a common prophylactic feature of any system of law.

Keeping these severe limitations in mind, the various kinds of evidence which the reported cases recognize as probative of the ultimate fact may be described in a number of categories.

A. STATISTICAL EVIDENCE. This kind of evidence is grounded on the mathematical theory of the law of large numbers, the fundamentals of which were developed by Laplace. In general, it may be said that statistical evidence so far dealt with by the cases falls in two classes. One consists of “demographic statistics” and the other of “comparative statistics.” The choice of terms is an unhappy one since, in discrimination cases, all uses of statistics are “comparative” in the sense that one set of numbers is inevitably matched up against some other set of numbers.

Second, it must be kept in mind that the calculation of probabilities a priori is based on the assumption that the outcome of each trial of an event in a series is governed by pure chance or “randomness”. Each toss of a perfectly balanced coin has an equal random chance of turning up heads or tails. Each roll of a perfectly formed and balanced die has an equal random chance of displaying at the top any number from one to six. Each spin of a roulette wheel, perfectly formed and balanced, and with a perfectly rounded and balanced ivory ball, has an equal random chance of having the ball come to rest in any number from one to thirty-six, on either a red or black number, or on either an odd or even number (disregarding for this purpose the two extra “house” spaces for zero and double-zero).

These a priori calculations of probability, being based on the assumption of pure chance, also imply that the outcome of one trial has no cause-and-effect relation to the outcome of another trial; if it did, the “pure chance” assumption would be destroyed. It is for this reason that heads can come up ten times in a row consistent with a priori probabilities, and the historical list of events in no way requires that the eleventh toss should come up tails. Each trial is independent.

The law of large numbers is designed to measure and articulate the probability that the outcome in a large number of trials will deviate, to one or another extent, from the a priori expectations, on the basis of the assumed “pure chance”.

Examined retrospectively (or “a posteriori"), the greater the deviation shown by the outcomes of a large number of trials in comparison with the a priori expectations, then the less is the probability that those results were of pure chance. But the law of large numbers does not purport to assert that a consecutive string of 1,000 heads is not the result of pure chance. It says no more than that such a series is extremely unlikely, puts a probability number on it, but also accepts that such a series can be the result of pure chance.

In applying these concepts to discrimination cases, it must be kept in mind that if the outcome of a series is so improbable as to warrant the inference that some influence other than pure chance is at work, the mathematics are of little use in arriving at a rational conclusion to identify what external influence is causing the deviation. To make a decision in that regard calls for some kind of evidence other than statistical evidence. If the subject is the tossing of coins or the rolling of dice, the kind of evidence looked for would be physical. Is the coin perfectly balanced statistically and dynamically? Is there an atmospheric movement that affects the spin? Is the surface on which the coin falls level, true and perfectly smooth? Is that surface soft enough to damp out any tendency to bounce?

When the subject is discrimination in employment, this kind of inquiry is considerably more complex and difficult. The precise physical data available for the study of coins and dice simply do not exist for interrelated human behavior. And, in addition, even if the proofs be clear that some influence other than pure chance is at work, the applicable law for a given claim only recognizes specific external influences as invidious or forbidden. If the external influence be one or more other than those denounced by the applicable statute, the abnormal outcome cannot of itself support the claim.

The demographic statistical approach compares two populations in respect to their mix and distribution of particular kinds of individuals. One population is the relevant segment of the particular employer’s work force. The other population is the group of individuals in the relevant labor market, and their mix and distribution. This kind of comparison of numbers resembles the process typical of some kinds of anti-trust cases, and a prime example of its application is that found in Int’l Brotherhood of Teamsters v. U. S., 431 U.S. 324, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977).

The second kind of statistical data used in discrimination cases is found where the claim is that some eligibility rule or practice arbitrarily and without rational basis automatically excludes some large segment. These are labeled as “disparate impact” cases, and a major example is Dothard v. Rawlinson, 433 U.S. 321, 97 S.Ct. 2720, 53 L.Ed.2d 786 (1977).

As noted above, both forms are really “comparative”, and the classification reflects no more than the differences in what composes the bases for comparison. The relevant work force, or applicants, or some other employer-related population is inevitably compared with a larger, non-employer related population. Sometimes, comparisons are made between the mix and distribution of one employer’s work force with that of another, as in Hazelwood etc., v. U. S., 433 U.S. 299, 97 S.Ct. 2736, 53 L.Ed.2d 768 (1977), decided the same day as Dothard.

And, although it is a jury selection case rather than an employment discrimination case, Casteneda v. Partida, 430 U.S. 482, 97 S.Ct. 1272, 51 L.Ed.2d 498 (1977) is worth careful reading for its discussion of the significance of statistical evidence, even though it appears to reflect an underlying distrust of the “key-man” system of jury selection more than confident rulings on the statistical data. Castaneda is also of value for what it does not say in discussing the significance of the outcomes of small samples (16 to 20 persons), each taken in 11 different years in a series, and adding them together as though they were a single “large” sample of 870 persons, and comparing the data with census figures of mix and distribution for the single year of 1970.

An obvious method by which an employer might employ statistics as a means for protection against discrimination claims would be to employ a system of stratified selection. This involves making a determination of the mix and distribution of the relevant labor pool, and filling vacancies with competent workers in proportion to that population. Such a course would tend to create a work force whose mix and distribution matches very closely that of the labor pool. The system would be similar to that of stratified random sampling widely used in the conduct of polls or other samplings. The difficulty is, however, that from a legal standpoint such an approach necessarily involves a deliberate purpose to select according to race, sex, etc., which are the very criteria forbidden by law.

The only instance so far reported where the Supreme Court has allowed a system even remotely resembling this approach is that in United Steelworkers v. Weber, 443 U.S. 193, 99 S.Ct. 2721. 61 L.Ed.2d 480 (1979), involving an “affirmative action” program at least in cases where the employer feels exposed to claims of “arguable violations” for past periods. In principle, the same approach ought to be valid in cases where there cannot be any arguable claim of past violations, e. g., where a new plant or facility is about to be established, in order to avoid future arguable claims, but the Supreme Court has not gone that far.

Another difficulty with stratified methods is that the mix and distribution of the relevant labor force are not static. They shift over time. Since most employers prefer a relatively stable work force, changes in the relevant labor market will throw the work force distribution out of kilter. Beyond that, a successful program of this kind would have a natural tendency to attract to the relevant market individuals who perceive themselves as being subjects of discrimination elsewhere. If this ensued, success at the start would generate failure in the future.

The paradox is not unlike the parallel one in the public schools. Once top-quality city schools have been confronted with heavy migrations from other States where segregated schools have failed to provide adequate education. Incoming pupils who have finished the 7th grade in a poor school, for example, may only be up to 5th grade level in a good school. The pupil cannot be placed in the 8th grade of the good school because he is not up to the work; nor can he be put in the 5th grade of that school because he would be “demoted” through no fault of his, and would be some three years older than his classmates. If a city school system were to invent a kind of upgrading school, to bring the incoming pupil to the level of his age group over a period of several years, and if it did so successfully, the likely result is that it would draw to itself more disadvantaged pupils from other areas than its capacity could handle, aside from shouldering the extra financial burden of curing deficiencies caused by inadequate school systems elsewhere. Here, too, success would invite failure.

Finally, the comparison of statistical data must be approached with care, especially when demographic mix and distribution of the relevant labor market is involved. Many reported cases suggest the view that if the mix and distribution of the employer’s work force does not reasonably match those of the relevant labor market, there is a strong inference that invidious influences other than chance are at work. The view would have some validity if the mix and distribution of the market were essentially stable over time, and if the employer’s workforce were wholly replaced periodically at fairly frequent intervals. Subject to the proofs in a given case, neither is likely to be true. With a highly mobile population, the mix and distribution of the market is more likely to shift rather than to be stable, and there will be even larger shifts among the individuals composing each segment of that market. At the same time, at least in most well-managed enterprises, the employer’s work force will likely tend to be stable in terms of the individuals composing it. When the employer succeeds in hiring, training and advancing individuals on merit, leaving everything else to chance, his work force will tend to be stable over long periods, with new hires coming in at entry level as long-term employees retire or die, and those between moving upwards as their skills and abilities improve. For this theoretically ideal model, it would be most unlikely that the work force mix and distribution would ever reasonably match those of the relevant labor market. It would tend to lag.

For this reason, such statistics should involve comparisons of the mix and distribution of the work force after “ageing” it (much as accounts receivable are “aged”), so that the aged groups can be compared to the market composition at corresponding times in the past.

For employers whose work force is represented by organized labor, the regular work force will usually enjoy a form of tenure and seniority rights (except for provisional workers). The employer is not free to create vacancies at various levels in order to hire others so that his workforce today can reasonably match the labor market. Even if unorganized, sound business reasons would fully justify keeping good, experienced and trained employees, and it would run counter to good business practice to discharge some of them to create job openings for new hires so that his mix and distribution will tend to match the current labor market.

For these reasons, demographic statistics will have the most meaning when compared to the mix and distribution of job applicants and of new hires, rather than to the mix and distribution of the entire embedded work force.

A final problem worth discussing is that of the sincere employer, who does not in fact consider any of the forbidden criteria, but who is subjectively perceived (incorrectly) by individuals in the labor market as being biased or prejudiced against them. For that situation, the employer can be hanged by the statistics even though to do so would be unjust.

No reported case discusses this kind of problem, but it seems obvious that the only course available in such a case would be one calculated to correct the falsely grounded perception of individuals in the market. One method that might be used, without violating any of the statutes, would be to adopt a hiring rule that when there are two applicants and only one job opening, and one of the applicants is from an untrusting group, then so long as the two are more or less equally qualified, the employer will hire both. A policy of this kind could, in time, constitute an actual performance that would correct the false perception. However, economic considerations being what they are, a program of this kind could only be used by an employer who has entry level job openings occur with reasonable regularity. In such a case, the employer would in effect be filling both the existing job opening and the next one anticipated that has not yet occurred but will. If the time period between job openings is reasonably short, such a program would not unduly burden an employer with more employees than he needs.

Obviously, such a program cannot be used to deal with promotions, especially where the structure of the work force is hierarchical, with fewer and fewer positions existing as one moves upward.

B. DISPARATE TREATMENT EVIDENCE. In cases where statistical evidence, whether demographic or comparative (i. e., to show the exclusionary impact of some facially neutral criterion), is either not available or does not tend to support the claim, the evidence tendered may be of the kind referred to as “disparate treatment” evidence.

In a statistical evidence case, given a disparity of some 2 or 3 standard deviations, or a “gross” exclusion of a significant percentage of otherwise potential appl