# Associated Gen. Contractors of America v. Columbus

> District Court, S.D. Ohio · August 26, 1996 · 936 F. Supp. 1363

URL: https://www.frixlaw.com/law-library/cases/1487035

## Case

- **Full name:** The ASSOCIATED GENERAL CONTRACTORS OF AMERICA, Et Al., Plaintiffs, v. CITY OF COLUMBUS, Et Al., Defendants
- **Court:** District Court, S.D. Ohio
- **Decided:** August 26, 1996
- **Citations:** 936 F. Supp. 1363; 1996 U.S. Dist. LEXIS 12519; 1996 WL 492336
- **Precedential status:** Published
- **Opinion:** Opinion by Graham
- **Judges:** Graham
- **Cited by:** 6 later opinions in the Frix Law Library

## Citator (automated)

- No negative treatment found by the automated citator. That is not the same as a confirmation that the case is good law; read the citing cases.
- Full citator and citing cases: https://www.frixlaw.com/law-library/cases/1487035

## Opinion text

OPINION AND ORDER
GRAHAM, District Judge.
I.
SUMMARY OF OPINION
In this action the plaintiffs, The Associated General Contractors of America, Central Ohio Division (“AGC”) and Patricia Sobiech, its Assistant Executive Director,
1
challenged the constitutionality of ordinances enacted by the city of Columbus, Ohio which required that firms owned by minorities and women receive a certain percentage of the dollar amount of subcontracts awarded on city construction projects each year. The AGC is an association of contractors engaged in the construction business in the city of Columbus whose members regularly seek construction contracts on city projects, both as prime contractors and subcontractors.
On January 23, 1989, the Supreme Court of the United States announced its opinion in
City of Richmond v. J.A. Croson Co.,
488 U.S. 469 , 109 S.Ct. 706 , 102 L.Ed.2d 864 (1989), a landmark civil rights case which held that municipal affirmative action programs which set aside a certain percentage of city contracts for minority-owned firms must be supported by firm evidence of past discrimination in city contracting. The Court struck down a program of the City of Richmond, Virginia which required, on an annual basis, that 30% of the city’s construction dollars be paid to minority subcontractors. The city had justified this program on the ground that the city’s population was 50% black. The Court held that the city should have focused instead on the number of qualified minority contractors and whether they had received a proportionate share of city construction dollars.
For many years, the city of Columbus had an affirmative action program for minority- and female-owned business enterprises (collectively “M/FBEs”) which included subcontracting goals of 10% for minority business enterprises (“MBEs”) and 2% for female business enterprises (“FBEs”) in city construction. In January of 1989, the city raised its subcontracting set-aside goals to 21% and 10% respectively. This was done without any evidence of past discrimination in city contracting and without any evidence of the number of qualified minority contractors or the amount of city construction dollars they had received. Several months later, this lawsuit was filed. In January of 1991, the city agreed that its set-aside program was unconstitutional and consented to an order which enjoined it from enacting any laws containing race- or gender-based preferences in city contracting without first obtaining the approval of this court.
In the aftermath of the
Croson
decision, the city immediately began efforts to find evidence of discrimination which would support new affirmative action legislation. It hired consultants and held public hearings. Finally, in December of 1993, the city enacted the Equal Business Opportunity Code of 1993 (“EBO Code”). This legislation provides a variety of race- and gender-based preferences in city contracting, including subcontracting goals of 10% for MBEs
2
and 7% for FBEs, as well as publicly funded bonding, financing and technical assistance programs for M/FBEs. In February of 1994, the city asked the court to dissolve the prior injunction and to permit the EBO Code to take effect. A trial was held to determine whether the new legislation and the evidence relied upon by the city meet the requirements of
Croson.
The city’s consultants collected data on the number of M/FBE construction firms in the Columbus Metropolitan Statistical Area (“MSA”) in order to calculate the percentage of available M/FBE firms. This is referred
*1372
to as the rate of availability. The city’s consultants also calculated the percentage of city contracting dollars that were paid to M/FBE construction firms. This is referred to as the rate of utilization.
With regard to the availability of M/FBE construction firms and their utilization as subcontractors on city construction projects, the best data presented by the city’s consultants showed the following:
Availability of M/FBE construction firms in the Columbus MSA as a percentage of all firms
MBEs FBEs
2.27% 4.49%
Utilization of M/FBE construction firms as a percentage of city subcontracting dollars
Year MBE Firms FBE Firms
1990 22.8% 10.5%
1991 8.2% 1.8%
1992 6.7% 13.0%
1993
3
5.7% .2%
This data shows that for each of the years studied, the MBE share of city subcontracting dollars exceeded their proportionate representation in the Columbus construction market. The share of women-owned firms fluctuated widely, but for the entire three and one-half-year period, their share averaged 9.2%.
The city pointed to the decline in MBE utilization after the suspension of the set-aside program as evidence of discrimination. The plaintiffs argued to the contrary that the decline in utilization of MBEs was attributable to the fact that the previous set-aside program had artificially inflated their share of subcontracting dollars to a level far above their rate of availability and that the decline simply represented the adjustment expected in a subcontracting market freed from set-aside requirements. Plaintiffs’ argument was supported by the city’s statistical evidence which showed that under the previous set-aside program, MBEs had received 33.5% of all subcontracting dollars in 1984, 47.6% in 1986 and 27.9% in 1988.
4
While the utilization of M/FBEs has decreased since the suspension of the previous set-aside program, their share of subcontracting dollars has nevertheless remained well above their rate of availability. Although the trial was held in May of 1995, the city did not provide the court with statistical evidence of the rate of M/FBE utilization after May of 1993, so the court is unable to determine whether the rate of utilization of M/FBE firms has continued to decline. The city’s statistical evidence does not support a finding of discrimination against M/FBE firms in city subcontracting.
In Columbus, prime contracts are awarded on the basis of competitive bidding and all contracts over $10,000 must be approved by city council and the mayor. There was no evidence that the city ever failed to award a prime contract to a minority firm that was the lowest bidder.
The city did not determine the number of M/FBE firms which are qualified to perform construction services for the city as prime contractors, so the evidence did not permit a comparative analysis of the prime contract dollars awarded to M/FBEs. However, there was evidence of the number of prime contracts awarded to M/FBEs which indicated that they received the following share: Year MBE Firms FBE Firms
1991
5
10.3% 1.3%
1992 7.48% 6.12%
1993 4.4% 4.44%
The city’s statistical evidence does not support a finding of discrimination in the award of prime contracts.
Under
Croson,
an affirmative action program may be justified even in the absence of evidence of discrimination by the municipality itself if there is firm evidence of discrimination in the private sector and that the city is an active or passive participant in that discrimination. When the city enacted the EBO Code of 1993, it relied on this ground also. In finding race and gender discrimination in the private sector of the Columbus construction industry, the city relied on statistical evidence which indicated that there
*1373
are proportionately fewer construction firms owned by women and minorities compared to the percentage of women and minorities in the general population of the Columbus metropolitan area and that the income of M/FBE firms is less than that of non-M/FBE firms. The evidence showed, however, that M/FBE firms tend to be smaller and newer than non-M/FBE firms, which would explain why their revenues are lower. The city’s evidence of the rates of business ownership was consistent with national demographic statistics. Comparative rates of business ownership are the result of many factors, including, no doubt, a history of discrimination against minorities and women in certain industries. However, in
Croson
the Supreme Court held that government programs which create race- and gender-based preferences cannot be justified by speculation about how many minority firms might exist absent past societal discrimination. 488 U.S. at 499 , 109 S.Ct. at 724-25 . As the Ninth Circuit Court of Appeals noted in
Coral Construction Co. v. King County,
941 F.2d 910, 925 (9th Cir.1991), “[t]he task of remedying society-wide discrimination rests exclusively with Congress.”
The city also investigated discrimination in city construction and the Columbus construction industry by interviewing minority and female business owners and by receiving testimony at public hearings. This kind of narrative evidence is generally referred to as anecdotal evidence.
The city’s investigation of anecdotal evidence of discrimination was poorly executed. It was not always focused on the relevant time and place. Many of the anecdotes were lacking in essential details. No efforts were made to verify reports of discrimination.' The emphasis of the investigation was on perceptions of discrimination, not actual discrimination. There was no attempt to determine whether similarly situated majority-owned firms were treated more favorably than M/FBE firms. Political pressures may have clouded the factfinding process. The city’s consultants, chosen without competitive bidding, were not impartial investigators, but aggressive advocates of minority set aside legislation as were some of the most vocal witnesses.
There was no evidence that complaints of discrimination had been filed against any of the city’s prime contractors or that there had been any findings of race or gender discrimination against local construction firms by the Ohio Civil Rights Commission or any court. There was no credible evidence that the city was guilty of race- or gender-based discrimination in the award of city construction contracts or that it was an active or passive participant in discrimination in the private sector.
Appellate courts which have interpreted and applied the
Croson
decision have held that anecdotal evidence alone will rarely suffice to show the kind of systemic pattern of discrimination necessary to support race- and gender-based preferences, and that anecdotal evidence is most useful as a supplement to strong statistical evidence.
See, O’Donnell Construction Co. v. District of Columbia,
963 F.2d 420, 427 (D,C.Cir.l992);
Coral Construction,
941 F.2d at 919 . Here, as noted, the statistical evidence did not support a finding of discrimination. The anecdotal evidence in this case fell far short of proof pf pervasive discrimination in the private sector.
Croson
requires that legislation containing race- and gender-based preferences must be narrowly tailored to accomplish the goal of remedying past discrimination. In order to satisfy this requirement, a local government must first consider race- and gender-neutral means for increasing M/FBE participation. In the construction industry, such race- and gender-neutral efforts would include programs which would provide all small or newly formed firms with technical assistance and assistance in bonding and financing. The city did not give serious consideration to race- and gender-neutral remedies before enacting the EBO Code. The EBO Code provides for technical assistance and assistance with bonding and financing, but although these programs will be paid for by the tax dollars of all citizens, their benefits are limited to firms owned by African Americans and women. Firms owned by other minorities, such as Hispanics, Native Americans and
*1374
Asian Americans are excluded, as are majority-owned firms.
Thirty M/FBE firms provided interviews and testimony about the Columbus construction industry. They were almost equally divided between the public sector and private sector of the construction market. This suggests that in Columbus, M/FBE firms do not rely heavily on governmental affirmative action programs, but that they are able to compete successfully in the private market. All of the M/FBE firms were experiencing some degree of success and many were quite successful. Half of those who responded to questions about firm income reported gross receipts in excess of $1 million per year.
There was evidence that M/FBE firms have occupied a prominent role in some of the city’s largest recent construction projects in both the public and private sectors, including the Arthur James Cancer Center, a $54 million-dollar project on which the Sherman R. Smoot Company was the general contractor; the much publicized Ohio Statehouse restoration project, a $112.7 million-dollar project, on which Smoot was the construction manager; the Tuttle Crossing mall, a $45 million-dollar project in which Moody/Nolan Ltd. is the project architect and the Smoot firm and Brothers Construction Co. are the general contractors; and the BaneOne Corporate Center at the Polaris Centers of Commerce, a $55 million-dollar project being built by a joint venture between Turner Construction Co., Brothers Construction Co. and Williams Builders. Smoot, Moody/Nolan, Brothers and Williams are all MBEs. The president of Brothers is an African-American woman.
The evidence showed that M/FBE set-aside programs do have some beneficial-effects. They require majority contractors to establish relationships with M/FBE firms and they give M/FBE firms the chance to demonstrate their competence. They provide an economic incentive for the formation of new M/FBE firms and provide opportunities for women and minorities in an industry in which they have been historically underrepresented. However, such programs may also have undesirable consequences. They may perpetuate racial and gender stereotypes by creating the impression that firms owned by minorities and women are not as capable as other firms. They may also breed racial and gender-based hostility when non-M/FBE firms lose work or are forced to pay higher prices because of them. Nevertheless, the court remains committed to uphold race- and gender-based preferences if they are supported by firm evidence of discrimination and they are narrowly tailored to accomplish a remedial purpose.
More than thirty years ago, Dr. Martin Luther King challenged our society to judge all individuals on the content of their character rather than the color of their skin. His dream was one of a colorblind society in which race no longer played a role in determining the success or failure of an individual. Today, in many respects, our society continues to suffer from the lingering effects of racial and gender-based discrimination. Notwithstanding this, the recent successes of African-Americans and women in every sector of our society suggests that there is movement, albeit gradual, toward the accomplishment of Dr. King’s dream. Clearly, our society has not fully achieved Dr. King’s dream, and this court does not assume that racism and sexism no longer exist, but there is evidence that significant progress has been made since Dr. King issued his challenge more than thirty years ago.
This court is bound to follow the law as set forth by the Supreme Court of the United States. That court clearly held in
Croson
that municipal programs which create race- and gender-based preferences violate the Equal Protection Clause of the Fourteenth Amendment of the United States Constitution unless there is a firm basis in the evidence that such programs are necessary to remedy past discrimination and that such programs are narrowly tailored to accomplish such a remedial purpose. After careful consideration, this court has found that the Columbus EBO Code of 1993 fails this test.
II.
HISTORY OF CITY’S AFFIRMATIVE ACTION LEGISLATION
The city of Columbus has had some form of affirmative action legislation affecting city
*1375
construction for over twenty years. The first such ordinance was enacted in May of 1975. It established an office of contract compliance and authorized the appointment of a contract compliance board. The affirmative action plan established by this legislation was directed at the racial composition of a contractor’s work force. It established a goal for the employment of minorities equal to the percentage of the minority population in the Columbus metropolitan area. Various enforcement mechanisms were provided. These provisions remained in effect until July of 1981 when the city added provisions which required construction prime contractors doing business with the city to award 10% of the value of their contracts to minority-owned firms and 2% to female-owned firms.
In May of 1983, the city created the Minority and Female Business Development Division (“MFBD Division”) to replace the office of contract compliance. The existing affirmative action goals were not changed. The MFBD Division was assigned the task of implementing and monitoring compliance with the city’s affirmative action ordinances.
The subcontracting goals of 10% and 2% remained in effect for seven and one-half years, until they were raised to 21% and 10% in January of 1989. The circumstances which led to the city’s decision to more than double the subcontracting goals were set forth in this court’s opinion in
Ohio Contractors Ass’n v. City of Columbus,
733 F.Supp. 1156 (S.D.Ohio 1990). Briefly, in 1990, the city was preparing to host an international floral and garden exposition known as Ameri-Flora ’92. The city’s black business community lobbied the city for a larger share of the financial benefits flowing from that project and persuaded it to insert 21% goals for MBEs and 10% goals for FBEs in the lease which controlled construction expenditures for that project. This led to the city’s decision on January 23, 1989, the same day
Cro-son
was decided, to amend its affirmative action ordinance by raising the minority and female set-aside goals to the same levels established for the AmeriFlora project.
In response to the Ohio Contractors Association’s challenge of the AmeriFlora affirmative action goals, the city repealed them and agreed to a permanent injunction eliminating them from all AmeriFlora projects funded by the city.
III.
THE EQUAL BUSINESS OPPORTUNITY CODE OF 1993
A.
Provisions of the EBO Code
The new affirmative action legislation which the city seeks to implement contains a variety of race- and gender-based preferences for M/FBE firms. The definition of minority is limited to African Americans.
Section 3921.01 of the EBO Code creates a twelve-member commission to be known as the Equal Business Opportunity Commission (“EBOC”) which has the responsibility for providing oversight of the implementation, review and modification of the Equal Business Opportunity Program.
Section 3921.02 of the EBO Code creates the Equal Business Opportunity Office (“EBOO”) as a division within the department of administrative services and provides that the EBOO administrator shall be its chief administrative officer with the responsibility for administration of the EBO Code.
Chapter 3922 provides for a system of contractor/vendor registration and M/FBE certifieátion which includes a graduation provision for firms whose annual sales exceed the average sales for firms in its sector of the industry for two consecutive years.
Section 3923.02 is captioned “Solicitation for Small Contracts” and directs the EBOO to establish procedures in conjunction with city agencies to enhance the award of contracts under $10,000 to M/FBEs by requiring a certain number of M/FBEs to be solicited on all requests for bids on contracts under $10,000. One of the procedures authorized by this section would involve soliciting bids from a total of three firms, two of which are required to be M/FBEs, with the third selected at random from a list of all registered vendors.
Section 3924.01 of the EBO Code requires the EBOO to establish a financing assistance program solely for the benefit of M/FBEs. This section directs the EBOO to examine
*1376
the feasibility of establishing “alternative and innovative” programs to assist M/FBEs in obtaining equity financing, and to examine the feasibility of developing a linked deposit program which would require the banks in which the city deposits its funds to establish comprehensive financing programs for M/FBE firms, including special lending programs. Section 3924.02(E) of the ordinance requires the EBOO to examine the feasibility “of using public funds to leverage private resources to establish a bonding pool for the issuance of bonds to M/FBEs on City contracts[.]” Section 3924.03 requires the EBOO to establish a clearinghouse of technical assistance programs and resources for M/FBEs. Non-M/FBEs are not entitled to participate in the benefits of the financing and technical assistance programs, nor would they be entitled to participate in the publicly financed bonding pool.
Section 3925.01 of the EBO Code requires the establishment of annual M/FBE participation goals for city contracting. Paragraph (B) of this section sets the initial annual M/FBE participation goals for construction at 10% for MBEs and 7% for FBEs. The ordinance states that the annual goals “are only intended to be benchmarks for evaluating the overall performance of the EBO program on an annual basis” and that they “are not, and shall not be quotas.” Section 3925.01(D). Paragraph (E) of this section provides: “On individual contracts or projects, there is no requirement that these annual M/FBE participation goals be met.” Paragraph (F) of this section provides that the city will make efforts to meet the annual percentage goals based • upon the dollar amounts of construction purchased from qualified M/FBEs and “[b]y establishing specific M/FBE goals on a contract by contract basis as set forth in Section 3926.01.”
Section 3926.01 requires the EBOO to “establish specific M/FBE goals on a contract by contract basis in order to achieve the objectives of this Code.” Specific contract goals are to be established according to criteria established by the EBOO which must include (but are not limited to) eight criteria listed in the ordinance. The specific M/FBE goals must be stated as part of the contract specifications put out for bid, and in order for a bidder to be deemed responsive, the bidder must either meet or exceed the stated M/FBE participation goals for the contract or demonstrate to the satisfaction of the EBOO that it has exercised good faith efforts to achieve the goals. The specific requirements of good faith efforts are listed in Section 3926.02, in the form of eight non-exhaustive factors. In addition, the EBO administrator may grant an administrative waiver of the M/FBE participation goals in accordance with Section 3926.03, which is captioned “Waiver for Detriment to Public Health, Safety and Welfare.”
Sections 3927.04 and 3927.05 authorize the EBOO to propose race- and gender-based preferences for M/FBE prime contractors, including price preferences and a sheltered market program subject to approval of city council and the mayor.
Section 3928.01 provides for administrative appeals from determinations of non-compliance with the requirements of the EBO Code or denials of certification as an M/FBE. Section 3928.02 provides criminal sanctions for fraud in seeking M/FBE certification, for false reporting of M/FBE utilization, and for “fronting” activity, which is defined as collusion between an M/FBE and a majority firm for the purpose of exploiting the M/FBE’s certification status to primarily benefit a majority firm without the performance of any commercially useful function by the M/FBE.
Section 3928.05 provides that the EBO program shall be terminated by city council and the mayor based upon a determination that the ongoing effects of marketplace discrimination in the Columbus MSA have been fully remedied and that statistical disparities in the utilization of M/FBEs have been eliminated in the public arid private sectors of the Columbus marketplace. Section 3928.06 is a sunset provision which requires that the EBO program shall be considered for reau-thorization five years from its effective date and may be extended for additional five-year periods upon a finding that its purposes and objectives have not been achieved.
*1377
B.
Effect of The EBO Code on Non-M/FBEs
Section 3926.01 of the ordinance creates barriers for non-M/FBE subcontractors which make it difficult or impossible for them to compete for a significant percentage of subcontracts on city construction projects. Although the EBOO is not required to adhere to the annual M/FBE participation goals for any single contract, it is required to achieve the EBO Code objectives, which include annual M/FBE participation goals amounting to 17% of total annual city construction contracting and procurement dollars. Thus, while a given- contract may or may not contain specific M/FBE goals, it is obviously the city’s intent to impose goals in sufficient amounts on a sufficient number of contracts to ensure that on an annual basis, M/FBE firms receive a percentage of total contract awards equal to the annual M/FBE goals.
A prime contractor who is not prepared to meet the M/FBE goals stipulated for a contract and who still wishes to bid must satisfy many additional requirements, and proceed with considerable uncertainty as to whether his bid will be deemed responsive. The practical effect of such a scheme is that a contractor bidding on a city contract which incorporates M/FBE contracting goals will set aside the stipulated percentage of the subcontract dollars for M/FBEs and make every effort to meet or exceed the M/FBE goal. Non-M/FBE subcontractors are unable to compete on an equal footing for all city contracting business, and if the annual goals are achieved, they will be effectively excluded from 17% of the city’s subcontracting market. Furthermore, non-M/FBEs are denied the benefit of bonding, insurance and technical assistance programs supported by public funds and public assets. Thus, the EBO Code of 1993 contains a variety of race- and gender-based preferences which must be examined in light of
Croson
and its progeny.
IV.
LEGAL PRINCIPLES APPLICABLE TO RACE- AND GENDER-BASED PREFERENCES
The Equal Protection Clause guarantees that no state shall “deny to any person within its jurisdiction the equal protection of the laws.” Fourteenth Amendment, § 1, United States Constitution. The central purpose of the Clause “is to prevent the States from purposefully discriminating between individuals on the basis of race.”
Shaw v. Reno,
509 U.S. 630, 642 , 113 S.Ct. 2816, 2824 , 125 L.Ed.2d 511 (1993). Thus, preferences based on racial criteria must receive a “ ‘most searching examination to make sure that they do not conflict with constitutional guarantees.’ ”
Wygant v. Jackson Bd. of Educ.,
476 U.S.. 267, 273-74, 106 S.Ct. 1842, 1847 , 90 L.Ed.2d 260 (1986) (plurality opinion) (quoting
Fullilove v. Klutznick,
448 U.S. 448, 491 , 100 S.Ct. 2758, 2781 , 65 L.Ed.2d 902 (1980) (opinion of Burger, C. J.)).
All governmental classifications by race are subject to “strict scrutiny,” regardless of whether they are supported by a “remedial” or “benign” purpose.
City of Richmond v. J.A. Croson Co.,
488 U.S. 469, 493-94 , 109 S.Ct. 706, 721-22 , 102 L.Ed.2d 854 (1989). In this circuit, gender-based affirmative action plans are also subject to strict scrutiny when challenged under the Equal Protection Clause.
Brunet v. City of Columbus,
1 F.3d 390, 404 (6th Cir.1993). The party defending the plan bears the burden of producing evidence that the plan is constitutional, while the party challenging the plan retains the ultimate burden of proving its unconstitutionally.
Aiken v. City of Memphis,
37 F.3d 1155, 1162 (6th Cir.1994).
The strict scrutiny analysis involves two inquiries: 1) Does the racial or gender classification serve a compelling governmental interest; and 2) are the means chosen by the governmental entity to protect that interest narrowly tailored to the achievement of that goal?
Adarand Constructors, Inc. v. Pena,
— U.S.-,-, 115 S.Ct. 2097, 2117 , 132 L.Ed.2d 158 (1995);
Aiken,
37 F.3d at 1162 . Strict scrutiny is meant to ensure that the purpose of a racial or gender preference is remedial.
Hopwood v. State of Texas,
78 F.3d 932 , 951 (5th Cir.1996). Requiring strict scrutiny is the best way to guarantee that courts will consistently give racial or gender classifications a detailed examination, both as to the ends to be met and the means used to achieve those
*1378
ends.
Adarand,
— U.S. at-, 115 S.Ct. at 2114 .
A government actor possesses a compelling state interest when its concern is remedying past discrimination.
Shaw,
509 U.S. at 654-56 , 113 S.Ct. at 2831 ;
United Black Firefighters Ass’n v. City of Akron,
976 F.2d 999, 1009 (6th Cir.1992). However, before classifications based on race or gender may be employed, the government actor must have “a strong basis in evidence” of past discrimination on the part of the governmental entity and the need for remedial action.
Miller v. Johnson,
— U.S.-,-, 115 S.Ct. 2475, 2491 , 132 L.Ed.2d 762 (1995).
See also, Croson,
488 U.S. at 500 , 109 S.Ct. at 725 (requiring “strong” or “convincing” evidence of past discrimination by that governmental unit). A prima facie case of intentional discrimination is sufficient to support a public employer’s affirmative action plan.
Brunet,
1 F.3d at 405 . A mere assertion by the governmental entity that the remedial action is required is not sufficient.
Miller ,
— U.S. at-, 115 S.Ct. at 2491.
The party seeking to implement a program remedying the present effects of past discrimination must produce evidence that the identified effects were caused by past discrimination, that the effects are of sufficient magnitude to justify the program, and that the program was adopted to remedy the identified present effects of past discrimination.
Hopwood,
78 F.3d at 952;
Podberesky v. Kirwan,
38 F.3d 147, 153 (4th Cir. 1994). As the Supreme Court noted in
Croson,
488 U.S. at 501 , 109 S.Ct. at 725-26 , “blind judicial deference to legislative or executive pronouncements of necessity has no place in equal protection analysis.”
See also, F. Buddie Contracting Co. v. City of Elyria, Ohio,
773 F.Supp. 1018, 1027 (N.D.Ohio 1991) (presumption of regularity afforded legislative acts does not apply to classifications suspect under the Equal Protection Clause; legislative body cannot rest upon a generalized assertion as to the classification’s relevance to the legislative body’s goals).
The existence of societal discrimination alone cannot support a racial or gender classification.
Aiken,
37 F.3d at 1162 . Likewise, nonracial factors such as deficiencies in working capital, inability to meet bonding requirements, unfamiliarity with bidding procedures and an inadequate track record do not suffice.
Croson,
488 U.S. at 498-99 , 109 S.Ct. at 724-25 . Rather, there must be some showing of prior discrimination by the governmental unit involved, either as an “active” or “passive” participant.
Croson,
488 U.S. at 492 , 109 S.Ct. at 721 ;
Wygant,
476 U.S. at 274, 276 , 106 S.Ct. at 1847, 1848 ;
Aiken,
37 F.3d at 1162 . Thus if the governmental unit can show “that it had essentially become a ‘passive participant’ in a system of racial exclusion practiced by elements of the local construction industry,” the governmental unit “could take affirmative steps to dismantle such a system.”
Croson,
488 U.S. at 492 , 109 S.Ct. at 721 . A government unit can become a “passive” participant by, for example, paying public contract dollars to prime contractors who discriminate in the award of subcontracts.
Concrete Works of Colorado v. City and County of Denver,
36 F.3d 1513, 1523 (10th Cir.1994).
See also, Coral Construction Co. v. King County,
941 F.2d 910, 916 (9th Cir.1991) (infusion of tax dollars into a discriminatory industry may be sufficient).
The types of evidence routinely presented to show the existence of a compelling interest include statistical and anecdotal evidence.
United Black Firefighters,
976 F.2d at 1010-1012 . The two types of evidence are frequently used in conjunction,
see Croson,
488 U.S. at 509 , 109 S.Ct. at 730 , and, in fact, courts have expressed concerns about a governmental entity relying exclusively on one or the other.
See Maryland Troopers Ass’n, Inc. v. Evans,
993 F.2d 1072, 1077 (4th Cir. 1993) (noting that race-conscious remedies upheld by the Supreme Court have been supported by both statistical and anecdotal evidence, and that “[i]nferring past discrimination from statistics alone assumes the most dubious of conclusions: that the true measure of racial equality is always to be found in numeric proportionality.”);
O’Donnell Construction Co. v. District of Columbia,
963 F.2d 420, 427 (D.C.Cir.1992) (anecdotal evidence may suffice to prove individual claims of discrimination but rarely can such evidence show a systemic pattern of discrimination);
Coral Construction,
941 F.2d at 919 .
Where gross statistical disparities exist, they alone in a proper case may
*1379
constitute prima facie proof of a pattern or practice of discrimination.
Croson,
488 U.S. at 501 , 109 S.Ct. at 725-26 . Appropriate statistical evidence setting forth a prima facie case of discrimination is sufficient to provide a strong basis in evidence to support an affirmative action plan.
Aiken,
37 F.3d at 1163 . Where there is a significant statistical disparity between the number of qualified minority contractors willing and able to perform a particular service and the number of such contractors actually engaged by the locality or the locality’s prime contractors, an inference of discriminatory exclusion could arise.
Croson,
488 U.S. at 509 , 109 S.Ct. at 730 . However, the existence of a gross disparity is not conclusive on the issue of discrimination because other factors unrelated to race may account for the disparity, and those opposing the affirmative action plan may present evidence to rebut the inference of discrimination produced by a gross statistical disparity.
United Black Firefighters,
976 F.2d at 1011 . Statistical evidence may be rebutted by a neutral explanation for the statistical disparities or by attacking the statistics themselves by showing that the statistics are flawed, by demonstrating that the disparities shown by the statistics are not significant or actionable, or by presenting contrasting statistical data.
Coral Construction,
941 F.2d at 921 .
Anecdotal evidence is most useful as a supplement to strong statistical evidence.
Concrete Works,
36 F.3d at 1520 ;
O’Donnell Construction,
963 F.2d at 427 . The presence or absence of complaints of discrimination is also relevant.
See Cone Corp. v. Hillsborough County,
908 F.2d 908, 916 (11th Cir. 1990) (noting evidence of complaints made to the county regarding discrimination by prime contractors);
Long v. City of Saginaw,
911 F.2d 1192, 1196 (6th Cir.1990) (noting that no complaints regarding discrimination in police hiring had been filed with any government agency).
The second branch of the strict scrutiny analysis is whether the program at issue is narrowly tailored. The Supreme Court in
Croson,
488 U.S. at 507-508 , 109 S.Ct. at 729-30 , identified four factors to be considered: 1) whether the city has first considered race-neutral means to increase minority participation, such as a race-neutral program of city financing for small firms, but found those programs to be ineffective; 2) the basis offered for the set-aside percentage selected; 3) whether the program provides for waivers of preference or other means of affording individualized treatment to contractors; and 4) whether the program applies only to minority businesses who operate in the geographic jurisdiction covered by the program. Other considerations include the flexibility and duration of the relief, that is, whether the program is appropriately limited such that it will not last longer than the discriminatory effects it is designed to eliminate, the relationship of numerical goals to the relevant labor market, and the impact of the relief on the rights of third parties.
Adarand,
— U.S. at-, 115 S.Ct. at 2118;
United States v. Paradise,
480 U.S. 149, 171 , 107 S.Ct. 1053, 1066-67 , 94 L.Ed.2d 203 (1987).
In the context of the construction industry, race-neutral means may include simplification of a city’s bidding procedures, a relaxation of bonding requirements and other barriers caused by bureaucratic inertia, training and financial aid for disadvantaged entrepreneurs of all races, and the enactment of laws prohibiting discrimination in the provision of credit or bonding by local suppliers and banks.
Croson,
488 U.S. at 510-511 , 109 S.Ct. at 730-31 . Strict scrutiny does not require the exhaustion of every possible alternative, however irrational, costly, unreasonable or unlikely to succeed.
Coral Construction,
941 F.2d at 923 . However, the Sixth Circuit cautioned in
Aiken,
37 F.3d at 1164 , that political pressures may prevent a city from utilizing racially neutral remedies.
When such pressures are present, racially neutral remedies often will remain untried, absent legal action by those persons adversely affected by the race-based relief. Under such circumstances, the courts must take special care as they engage in their “most searching examination” of whether racial preferences have been shown to be necessary,
Wygant,
476 U.S. at 273 , 106 S.Ct. at 1846 , since that examination will not be undertaken by any other body.
Id.
The scope of the remedy must depend' upon the scope of the violation.
*1380
O’Donnell Construction,
963 F.2d at 425 . To pass constitutional muster, racial or gender classifications must be necessary to the accomplishment of their legitimate purposes.
Long,
911 F.2d at 1196 . Limiting the duration of a race- or gender-conscious remedy “which clearly impacts adversely upon the plaintiffs is a keystone of a narrowly tailored plan.”
Detroit Police Officers Ass’n v. Young,
989 F.2d 225, 228 (6th Cir.1993).
Flexibility forbids the mechanistic application of fixed quotas.
Peightal v. Metropolitan Dade County,
26 F.3d 1545, 1557 (11th Cir.1994). A valid program should include a waiver system that accounts for the availability of qualified minority or female businesses and whether those businesses have suffered from the effects of past discrimination by the governmental unit or prime contractors operating within the governmental unit’s jurisdiction.
Coral Construction,
941 F.2d at 924-925 .
See also, Croson,
488 U.S. at 508 , 109 S.Ct. at 729-30 (noting the defects in a waiver provision which-focused solely on the availability of minority business with no inquiry into whether or not the particular minority business seeking a racial preference had suffered from the effects of past discrimination by the city or prime contractors.) A valid program must also be limited in its effective scope to the boundaries of the enacting jurisdiction.
Coral Construction,
941 F.2d at 922 . The relevant question is not one of business location, but of business participation in the construction industry within the relevant geographical area, that is, whether the business has been discriminated against within the boundaries of the governmental entity.
Id.
at 925 .
V.
THE CITY’S EFFORTS TO ESTABLISH A FACTUAL PREDICATE FOR RACE- AND GENDER-BASED PREFERENCES
A.
The City’s Outside Consultants
1.
Beatty & Roseboro (1989)
The city’s efforts to establish a factual predicate for race- and gender-based' preferences in city contracting began in July of 1989. The court made reference to this initial effort in its previous order of June 20, 1990, pp. 10-15, which denied defendants’ motion to dismiss on grounds of standing and mootness:
The evidence offered by the plaintiffs in opposition to defendants’ motion to dismiss suggests that serious questions may exist regarding the impartiality and methodology of the procedures the City has followed in an effort to insure that its affirmative action ordinances comply with the constitutional mandate of the Supreme Court in
Croson.
As noted above, on July 17,1989, the City authorized the retention of outside consultants to conduct a statistical analysis of the City’s contracting history. The outside consultants were the law firm of Beatty & Roseboro, headed by Otto Beatty, Jr., a prominent black attorney and member of the Ohio General Assembly, who has sponsored several set aside bills. Mr. Beatty had offered his services to Columbus City Council President, Jerry Hammond, on June 5, 1989, indicating that he had “prepared a methodology to document the need for [a set aside plan] in the city.” Beatty’s correspondence indicates that both he and Hammond had strong preconceived notions regarding the need for a set aside program which presumed the existence of racial discrimination.
It was apparently Mr. Beatty’s advice which prompted City Council to amend its affirmative action ordinances by deleting the numerical goals for the specific purpose of rendering the issues in this case moot. Mr. Beatty’s legal strategy is set forth in his letter of November 21, 1989 to the Minority Business Enterprise Legal Defense and Education fund in Washington, D.C., in which he referred to his representation of both the City of Columbus and Montgomery County, Ohio and set forth his legal strategy ...
On November 20, 1989, Mr. Beatty issued a report entitled the “City of Columbus MBE/FBE Study” finding that during the period 1969 to 1989 minority business
*1381
enterprises received only 1.52% of the 1.2 billion dollars in construction spending by the city during that period. Following the report, Mr. Beatty conducted two public hearings on behalf of council on January 18th and March 8th, 1990.
Plaintiffs allege that the Beatty study overstates the monetary amounts expended on city contracts and understates the number of such dollars which flowed to MBEs.... Plaintiffs further allege that the Beatty report reflects only city monies paid directly to minority contractors by the city and fails to account for millions of dollars of payments by majority contractors to minority subcontractors on city projects .... Plaintiffs further point out that the Beatty study does not attempt to identify the pool of qualified minority business enterprises, which is a necessary predicate to a determination of whether or not a statistical analysis of contract dollars awarded to minority business enterprises is disproportionately low.
See Croson,
[ 488 U.S. at 469, 495-501, 507-08 ], 109 S.Ct. at 706, 723-25, 729 .
Plaintiffs raise further objections to the fairness and objectivity of Mr. Beatty’s fact finding process, specifically the use of his own “copyrighted” affidavit which was sent to over 1,000 MBEs over the signature of Council President Hammond. Plaintiffs suggest that this affidavit entitled “Data Discrimination Bank Form” was not a neutral questionnaire designed to determine whether discrimination had occurred but was instead “a blatant effort to manufacture evidence” because, by its very terms, it presumed the existence of discrimination and proffered alternative responses consistent with that assumption. (AGC’s memorandum contra motion to dismiss, p. 31). Plaintiffs also dispute the fairness of the hearings conducted by Mr. Beatty and refer to an exhibit consisting of a memorandum to Mr. Beatty from his legal associate describing the testimony of one of the witnesses at the January 18th hearing as follows:
The next speaker was a Mr. Randall Gaddis, who is the Second Vice President of Gaddis & Son, a major concrete contractor in Columbus. Mr. Gaddis stated that in March his company will have been in business in central Ohio for nineteen (19) years. Mr. Gaddis stated that there was no obvious discrimination which his firm had suffered. Mr. Gaddis is obviously a very important witness. His testimony is no good and in fact hurts the cause of set-asides. Specific instances of discrimination are needed from him. It is important that Mr. Gad-dis’ father be brought in, read the Riot Act and be made to submit an affidavit which would contradict that to which his son testified. (Benton depo, Ex. 16).
Plaintiffs suggest that Mr. Beatty has a deep personal and philosophical commitment to minority set aside programs and that instead of conducting an impartial factual investigation to determine whether or not there is a historical pattern of discrimination against minorities in employment, contracting and subcontracting for city construction, Mr. Beatty is operating as an advocate, marshalling and molding the evidence so as to justify a minority set aside program. Plaintiffs suggest that it is unrealistic for City Council to expect Mr. Beatty to protect the interests of majority citizens and firms as required by the Supreme Court’s
Croson
decision.
The court is not passing on any of these claims at this time. However, it does appear that plaintiffs have raised serious questions about the City’s intentions .and the propriety of its procedures in preparation for the reenactment of affirmative action goals which demonstrate that the issues in this case are not moot.
In light of the criticisms of Mr. Beatty’s study, the city decided to find another consultant to assist it in developing a factual predicate for affirmative action legislation.
2.
MBELDEF and BBC, Inc.
Council President Hammond assigned the task of finding a new consultant to council member Ben Espy, a prominent black lawyer and politician. Mr. Espy contacted the Minority Business Enterprise Legal Defense and Education Fund (“MBELDEF”), the same organization Mr. Beatty had corresponded with and confided his legal strategy
*1382
to in November of 1989. MBELDEF, located in Washington, D.C., describes itself as a “national nonprofit public interest law firm and membership advocacy organization founded in 1980 by former U.S. Congressman Parren J. Mitchell (D-Md.) for the purpose of providing legal defense of class interests of minority business enterprises ...” Thus, the city chose as its second consultant a firm that was in the business of advocating the interests of minority businesses.
MBELDEF was not qualified to conduct a statistical analysis of possible race and gender discrimination in city contracting and arranged for the Denver firm of Browne, Bortz & Coddington, Inc. (“BBC”) to do this part of the proposed disparity study. In May of 1991, the city accepted the MBEL-DEF/BBC proposal.
3.
Beatty & Roseboro and Bradford (1991)
In the meantime, Mr. Beatty had continued Ms work. According to the testimony of Council President Lazarus, he did so without the city’s knowledge or approval. Nevertheless, in January of 1991, Mr. Beatty presented the city with a pair of reports. The first, authored by his firm, was entitled “Columbus, OMo Disparity Study and Recommendation” (“Beatty Report”), and the second, which he commissioned from William D. Bradford, Ph.D., was entitled “Disparity in Public Contracting, Columbus, Ohio Economic Evidence and Recommendation” (“Bradford Report”). Although the city had decided to retain another consultant because of the issues raised eoncermng the objectivity and reliability of Mr. Beatty’s work, and even though it had not authorized him to do any further work, the city accepted the Beatty and Bradford reports and expressly relied on them when it adopted the 1993 EBO legislation.
B.
The Factual Predicate for the EBO Code of 1993
The disparity study commissioned from MBELDEF and BBC was completed in August of 1992 and was presented to the city in form of a report entitled “Predicate Study, City of Columbus, Ohio” (“Predicate Study” or “PS”). After receiving the Predicate Study, city council proceeded to hold public hearings on October 28 and 29 and November 18 and 19, 1992. These hearings were coordinated by MBELDEF. Thereafter, the city commissioned several additional studies from BBC, some of which were completed and received before the enactment of the EBO Code and some after. These additional reports and studies included the Predicate Study Supplement (“Predicate Study Supplement” or “PSS”) (September, 1993); the Management Study (March, 1993); the Employment Predicate Study (“Employment Study” or “ES”) (September, 1994) and the Predicate Study Second Supplement (“Predicate Study Second Supplement” or “PS 2d Supp.”) (January, 1995).
When the city enacted the EBO Code of 1993, it recited its reliance
inter alia
on the public hearings held in January and March, 1990 and October and November, 1992, the Beatty Study, the Bradford Study, the Predicate Study, and the Predicate Study Supplement. In January of 1995, the city amended the ordinance to recite reliance on the Employment Study and the Predicate Study Second Supplement.
C.
Postenactment Evidence
At trial, the city offered into evidence the Management Study, the Employment Study and the Predicate Study Second Supplement even though the city did not rely on them when it enacted the EBO legislation. Indeed, the last two were not in existence when the legislation was enacted.
Most courts wMch have been presented with the question of the admissibility of post-enactment evidence have held that it is admissible, particularly where the plaintiffs claim, as here, is for injunctive relief.
See, e.g., Concrete Works,
36 F.3d at 1521 (posten-actment evidence admissible if useful for court’s determination of whether ordinance’s deviation from norm of equal treatment was necessary);
Contractors Ass’n of Eastern Pennsylvania, Inc. v. City of Philadelphia,
6 F.3d 990, 1004 (3d Cir.1993) (consideration of postenactment evidence appropriate where principal relief sought and only relief granted was injunction);
Coral Construction,
941
*1383
F.2d at 920 (factual predicate for program should be evaluated based upon all evidence presented to district court, including evidence adduced after enactment). The court in
Concrete Works of Colorado, Inc. v. City and County of Denver,
823 F.Supp. 821, 887 (D.Colo.1993) commented that “it would make little sense to strike down” an ordinance for lack of factual support “only to watch the City Council reconvene immediately, incorporate the new evidence into a new ordinance, and arrive at a constitutionally adequate factual predicate.” The court in
Coral Construction,
941 F.2d at 920-921 , reasoned that this rule would relieve municipalities presented with some evidence of discrimination of the dilemma of having to choose between waiting until further development of the record, thereby risking liability to minorities, and acting prematurely, thereby risking liability to nonminorities.
The reasoning of
Coral Construction
would not apply to the instant case, where the city had an M/FBE set-aside program in effect for many years prior to the contested legislation which had resulted in substantial overu-tilization of MBEs.
See, supra,
p. 1372. In Columbus, there would appear to be little or no risk that a delay in adopting race- and gender-based preferences would involve a risk of liability to minorities.
Courts have also noted potential disadvantages to using postenaetment evidence.
See Contractors Ass’n of Eastern Pennsylvania,
6 F.3d at 1004 (recognizing the risk of insincerity associated with postenaetment evidence);
Concrete Works,
823 F.Supp. at 837 (noting the plaintiffs opposition to consideration of postenaetment evidence based on concern that it is not possible to ensure that program is remedial if cause for' remedy is not discovered until ordinance takes effect “has considerable merit”).
The admission of postenaetment evidence poses a risk of other undesirable consequences. It may encourage a government which has a strong political motivation to enact race- and gender-based preferences to proceed without an adequate factual basis, gambling that the legislation will not be challenged in court, and knowing that further efforts can be made to find a basis for the legislation while litigation is pending. The rule likewise discourages citizens who are adversely affected by such legislation from mounting a constitutional challenge even when good grounds exist, knowing that they face a moving target and that once challenged, the city may spend enormous sums to marshal evidence to support the legislation. Thus, the chilling effect of the rule may leave unchallenged race- and gender-based preferences which have no remedial justification. This court would hold the postenaetment evidence inadmissible. However, in this opinion, the court will consider all of the evidence which the city has offered in support of the legislation.
VI.
CITY’S PROCUREMENT PROCEDURES FOR CONSTRUCTION SERVICES
The expenditure of city funds is governed by the Columbus City Charter and Chapter 329 of the Columbus City Code. All purchases of goods, materials and services are subject to competitive sealed bidding. Contracts must be awarded to the lowest responsible and responsive bidder. A responsible bidder is defined as one “who has the capability, capacity, facilities, equipment and personnel needed to fully perform the contract requirements, and the experience, integrity, reliability and credit which will assure good faith performance of the contract requirements.... ” Columbus City Code § 329.04(1). A responsive bidder is one who has submitted a bid which conforms in all material respects to the requirements set forth in the invitation to bid. Columbus City Code § 329.04(m).
The individual city agency or department undertaking a construction project must prepare an invitation to bid which contains specifications and contract terms and conditions. The Division of Public Service customarily includes the following requirements in its invitation to bid:
The Bidder is required to state, in detail, in the space provided below, what work of a character similar to that included in the proposed contract he has done, to give references and such other detailed information as will enable the Director of Public
*1384
Service to judge of Ms responsibility, experience, skill and financial standing. Bids from Contractors inexperienced in tMs particular work will not be considered. Among other things, tMs statement shall include the following: Evidence to the effect that the Bidder maintains a permanent place of business; list of plant equipment available for the Work under the proposed contract; together with the statements as to when purchased or otherwise obtained and statements as to its present physical condition; evidence to the effect that the Bidder has a suitable financial status to meet obligations incident to the Work, and evidence to the effect that the Bidder has appropriate techmcal experience.
Invitations to bid are sent to a list of construction firms selected by the agency. In some cases, firms receiving solicitations may be prequalified by the agency. All invitations to bid must be advertised in the Columbus City Bulletm which is available to anyone for a subscription fee of $30 per year. They may also be advertised in newspapers, trade journals and other appropriate publications. They are also posted in the reception area of the office of the City Purchasing Division. Bids must be submitted in writmg. When all of the bids are received, they are reviewed for responsiveness. Every bid must be accompamed by a bid bond and a performance bond of at least 10% and 50%, respectively. In practice, the city customarily requires a 100% performance bond on all contracts for public improvements.
The bids are opened and publicly read at the place, date and time specified in the notice in the presence of the city auditor or his representative. Local bidders, i.e., those located withrn Franklin County, receive a credit equal to 1% of the lowest bid submitted by a non-local bidder. When the wiiming bidder has been selected, the department •then prepares legislation for approval by the city council and the mayor. Whenever a contract is awarded to any bidder other than the lowest, the director of the city agency must provide a written explanation to city council. The contract must also be reviewed and approved by the city auditor and the city attorney. No contract is effective until it is approved by an ordmance of city council.
For purchases over $500 but under $10,-000, invitations to bid are posted but not advertised. Typically a city department seeking construction services on a contract under $10,000 will contact three potential vendors for quotes. The local preference credit for these purchases is 5%. Although the procedures for such contracts are simplified and do not require approval by council and the mayor, nevertheless, the contract must be awarded to the lowest responsible and responsive bidder. Biddmg on these contracts is not limited to those firms invited to bid but is open to any firm who wishes to bid. Anyone wishing to bid on such projects would have to regularly inspect the public notices posted in the offices of the Purchasing Division.
Until April, 1991, advertising of invitations to bid and approval by council and the mayor was required for all contracts in excess of $5,000.
The above procedures apply to prime contracts. Prime contractors themselves determine what procedures they wish to follow in awarding subcontracts. Normally, a prime contractor would privately solicit bids from subcontractors of its choice. A prime contractor may or may not require a subcontractor to furmsh a performance bond.
VII.
EVIDENCE OF DISCRIMINATION IN CITY CONSTRUCTION
The evidence of discrimination against M/FBEs which the city relied upon in enact-mg the EBO Code included both statistical and anecdotal evidence. The investigation mcluded both city construction and the private sector of the construction industry in the Columbus MSA. The court will first analyze the statistical and anecdotal evidence of discrimination in city contracting and will then consider both kinds of evidence of discrimination in the private sector.
A.
Statistical Evidence of Discrimination in City Construction
1.
Beatty Report
Beatty & Roseboro undertook a statistical study to investigate discrimination in the
*1385
award of city contracts. The Beatty Report is included in the record as Ex. D272. Beatty & Roseboro purportedly determined the percentage of total city expenditures made to MTFBEs for goods, services and construction and compared it to the percentage of eligible MTFBEs in the relevant markets. They referred to this comparison as the utilization/eligibility index.
Beatty & Roseboro failed to establish that any of the M/FBEs included their definition of “eligible” were in fact qualified and able to provide construction services to the city. At page 38 of their report, Beatty & Roseboro concede that this is a problem. Instead of attempting to determine whether a firm was qualified to perform construction services for the city, they chose to define “eligible” as “any firm that is considered to be in business.” There is no basis for the assumption that all construction firms in business are capable of performing construction services for the city.
Beatty & Roseboro limited their economic analysis to prime contracts. Accordingly, their list of eligible firms should have been limited to firms which were capable of qualifying as prime contractors. City ordinances require bid bonds and performance bonds on all contracts for public improvements. Clearly not every firm “in business” can meet the financial requirements to obtain such bonds. The failure to establish that “eligible” MTFBE firms were qualified to provide construction services to the city as prime contractors deprives the Beatty study of any probative value.
In order to determine the total number of all available construction firms for the pirn-pose of computing utilization ratios, Beatty
&
Roseboro used statistics for Franklin County for the years 1969 to 1987 published in
County Business Patterns (Ohio),
a publication of the U.S. Department of Commerce, Bureau of the Census. Beatty & Roseboro arrived at MBE eligibility ratios for construction ranging from 5% in 1980 to 10% in 1989, and from 1% to 2% for the same years for FBEs.
A review of Beatty & Roseboro’s list of eligible M/FBEs shows that it includes trucking companies, landscape and horticultural firms and building maintenance firms. Such firms account for almost 20% of the total number of firms listed as available MBE construction firms.
County Business Patterns
does not include trucking companies, landscape and horticultural service companies or building maintenance firms in the construction category. By including these businesses in calculating the number of available MTFBE construction firms while using data for the total number of construction firms which excludes them, Beatty & Rose-boro overstated the percentage of available M/FBE construction firms.
To determine M/FBE utilization, Beatty & Roseboro relied on information prepared by the city auditor’s office. Their report states that they gave the auditor the vendor numbers of all available M/FBEs and asked him to prepare a report of all payments to those firms for construction services.. Construction services were identified by codes which the auditor uses in his computerized accounting system. The resulting total of payments to M/FBEs was compared to the total payments for construction services to all firms. Using the auditor’s accounts payable data on prime contract dollars and their list of M/FBEs, Beatty & Roseboro calculated utilization ratios for MBEs ranging from .4% in 1980 to. 2.5% in 1989, with a high of 6.5% in 1984. For FBEs, the range was 0% in 1980 to .0002% in 1989, with a high of 1.28% in 1984.
This data is limited to payments to prime contractors and ignores all of the city construction dollars that were paid to minority subcontractors. By failing to consider the amount of dollars which were paid to M/FBE subcontractors, the Beatty Report omits information which is vital to a determination of whether M/FBE construction contractors received a disproportionately low percentage of total city construction contract dollars. This is precisely the same omission which contributed to the downfall of the Richmond plan in
Croson. See
488 U.S. at 502 , 109 S.Ct. at 726 .
The report indicates that Beatty & Rose-boro were provided with data compiled by the city’s MFBD Division which showed the dollar amounts of subcontracts awarded to M/FBEs. Beatty & Roseboro noted that
*1386
this data was based on contracts awarded instead of payments made and that the city did not audit the reports submitted by prime contractors. Thus, Beatty & Roseboro concluded that this data was unsubstantiated and they declined to use it. They did acknowledge, however, that they were able to verify payments made on many of the contract awards. Although Beatty & Roseboro may have been justified in refusing to accept the M/FBE Division’s data at face value, they failed to adequately explain why they did not take steps to verify it. By contrast, they apparently spent hundreds of hours sifting through records of payments for goods and services in order to compile utilization data for those markets. Beatty Report, p. 40. Their skepticism about' the subcontracting data was not shared by the city’s subsequently retained statistical consultant, BBC. After taking steps to verify this subcontracting data, BBC concluded that it resulted in “a highly accurate database of prime/subeon-tract participation.” PSS, p. 2. In any event, Beatty & Roseboro’s skepticism about the MFBD Division’s subcontracting data does not justify their decision to completely ignore subcontracting in investigating whether or not discrimination existed in the award of city construction contracts. After all, the proposed legislation creates race- and gender-based preferences in subcontracting.
At pages 38 and 39 of the report, Beatty & Roseboro indicate that they initially considered using the
Survey of Minority-Owned Business Enterprises
and
Women-Owned Businesses,
published by the U.S. Department of Commerce, Bureau of the Census, to determine the number of eligible M/FBEs, but rejected that option. Nevertheless, beginning at page 47 of the report, without any farther explanation, they undertake an analysis of data from these sources, stating that “when the data is compared with
County Business Patterns Ohio
census data, a new Eligibility Ratio is the result.” These calculations are presented in numerical and bar-graph form, beginning at page 48 of the report. Here, the report presents new eligibility ratios of twelve for MBEs in construction for the years 1982 and 1987. Beatty & Roseboro do not explain how this ratio was calculated, but it appears that they used the
Survey of Minority Owned-Business Enterprises
and
Women-Oumed Businesses
to determine the total number of available M/FBEs and
County Business Patterns (Ohio)
to determine the total number of firms. This methodology mixes apples and oranges because the
Survey of Minority-Owned Business Enterprises
and
Women-Owned Businesses
include minority- and women-owned firms from a much larger geographical area than
County Business Patterns.
The former publications include firms in the seven-county Columbus MSA, while the latter is limited to Franklin County.
The Beatty Report does not provide a basis for a finding of race or gender discrimination in the award of city contracts. The statistical analysis was flawed by its failure to consider subcontracting, and the analysis of prime contracting awards was flawed by the failure to establish how many M/FBEs were qualified to bid on prime contracts. Beatty & Roseboro’s attempt to use census data to show underutilization of M/FBEs in city construction was further flawed by their error in combining two different kinds of census data which were based on different populations of construction firms.
2.
The Bradford Report
Beatty & Roseboro retained William D. Bradford, Ph.D., to conduct a statistical analysis of the economic evidence contained in their report. Dr. Bradford calculated the standard deviation for the binomial distribution between the amount of prime contract dollars M/FBEs would have been expected to receive based on the availability ratio calculated by Beatty & Roseboro and the actual amount they received as reported by Beatty & Roseboro. Using three standard deviations as the test of statistical significance for a prima facie finding of discrimination, Bradford reported that the apparent discrepancies were statistically significant. He then calculated, for each of the years of the Beatty
&
Roseboro study, the difference in dollar amounts and percentages between M/FBEs actual payments and their expected payments after deducting three standard deviations.
*1387
Bradford relied entirely on the availability and utilization ratios calculated by Beatty & Roseboro and on the economic data contained in their report. He did not verify any of their research or calculations. Bradford concluded his report with an expression of concern about the propriety of the statistical methodology he used, saying “The two or three standard deviation goal was set forth by the Court.” Bradford Report, pp. 8-9. Apparently Bradford used the standard deviation test of statistical significance because he believed the Supreme Court had endorsed it in
Croson.
He was mistaken.
See infra,
p. 1401.
Since Bradford’s report is based entirely on the invalid and incomplete data contained in the Beatty Report, it does not support a finding that the city discriminated against MTFBEs in the award of construction contracts.
3.
The Predicate Study
In August of 1992, BBC and MBELDEF presented the city with their 182-page Predicate Study. The Predicate Study contains a presentation of the statistical evidence collected by BBC and the anecdotal evidence collected by MBELDEF.
BBC began its statistical analysis of the Columbus construction market by selecting the Columbus MSA as the relevant geographic market area. The Columbus MSA is a functional economic area defined by the U.S. Department of Commerce which consists of the counties of Delaware, Franklin, Fairfield, Licking, Madison, Pickaway and Union. The city of Columbus Purchasing Division’s database of firms that have submitted bidder information revealed that in 1991, 62% of the firms were within the Columbus MSA and another 26% were from other Ohio communities. BBC determined that the chronological framework for its statistical analysis of city contracting would encompass the period January, 1983 through June, 1991. This period was selected because of the availability of computerized payment data.
a)
Utilization Data
BBC identified a number of different sources of purchasing information but decided to use the city auditor’s vendor payment history file. This database contains a record of every check issued to a business which provided goods or services to the city.
The vendor payment history file does not include information regarding the race or gender of the payee. Thus, it was necessary for BBC to construct a list of MTFBE firms and to cross-reference that list with the names of firms receiving payments recorded in the vendor payment history file. BBC could then calculate the total dollar amounts paid to M/FBEs.
The auditor’s vendor payment history file identifies the type of expenditure by minor object codes (“MOCs”). By selecting the MOCs that appeared to be related to construction, BBC generated data on estimated expenditures for construction services. However, the classification of MOCs into specific procurement areas and categories was only approximate. PS 1-13.
The vendor payment history file is the same source of data used by Beatty & Rose-boro. It only contains information on payments to prime contractors. A substantial amount of all dollars paid to prime contractors are used to pay subcontractors who may be of a different race or gender than the prime contractor. BBC acknowledged at PS II-7 that sole reliance on the auditor’s vendor payment history file might present an inaccurate picture of participation of minority- and female-owned firms.
BBC identified a source of data regarding subcontracting, namely, a subcontractor participation report that prime contractors were required to submit to the city’s MFBD Division. These reports provide information regarding the name of each subcontractor, the subcontractor’s MTFBE status and the dollar value of its subcontract. This data is based on reports submitted prior to commencement of a project. The MFBD Division staff entered data from the reports into a computer database. The same data was tabulated by the city’s Legislative Research Office. In order to ensure the integrity of this
*1388
data, BBC conducted a completely independent compilation of subcontractor participation and verified its data with the city departments which awarded the construction contracts. Through this process, BBC constructed its own independent database of information on prime and subcontract awards. Thus, BBC had two sources of data regarding city construction purchases: the vendor payment history file which contained information only on prime contracts, and the contract document database which contained information on both prime and subcontracts.
Citing the amount of work necessary to develop the information, BBC compiled subcontract data only for selected years: 1984, 1986, 1988, 1990 and January through June, 1991. There were large fluctuations in the amount and percentage of payments made to M/FBE subcontractors from year to year.
6
The lack of data for half of the years included in the study period precludes a complete analysis of the participation of M/FBE construction firms as subcontractors during the study period.
Using the vendor payment history file, BBC calculated the percentage of prime contract dollars paid to M/FBE firms for construction services each year for the years 1983 through June, 1991.
See
PS Table IIA-8. For the entire eight and one-half-year period, 2.2% of the total prime contract dollars went to MBEs and 1.3% went to FBEs.
7
The percentage paid to MBE firms ranged from a high of 2.8% in 1990 to a low of .9% in 1985. Payments to FBE firms ranged from a high of 7.5% in 1984 to a low of .2% in 1987.
BBC’s analysis of the subcontracting data revealed that M/FBE participation as subcontractors was much higher than their participation as prime contractors.
See
PS Table II-A-12. Indeed, for the years 1984, 1986 and 1988, MBEs received 33.5%, 47.6% and 27.9% of the subcontract dollars. The percentage of subcontract payments to FBE firms ranged from a high of 10.5% in 1990 to a low of 2.1% in 1988.
BBC’s contract document database contained information on prime contract awards, as well as subcontract awards. The data in the contract document database indicated a significantly higher utilization of M/FBEs as prime contractors than the data from the vendor payment history file.
Compare
PS Table II-A-9 with PS Table II-A-7. The contract document database indicated that MBE firms received 2.6% of the prime contracts in 1984, 1.8% in 1986, 4.2% in 1988, 7.7% in 1990 and 3.7% in the first six months of 1991.
See,
PS Table II-A-10. While the data in the contract document database allocated the dollars to the years the contract was signed and the vendor payment history file allocated the dollars to the years they were paid, this would not seem to account for the magnitude of the apparent discrepancies. In some years the prime contract dollars received by MBEs based on the contract document database are twice or more the amounts indicated by the vendor payment data. In the case of FBEs, the contract document database shows they received a smaller percentage of prime contract dollars in the earlier years of the study, but in 1990 and the first six months of 1991, the share of FBEs of prime contracts is shown to be 1.6% and 1.9% instead of the .8% and .6% shown by the vendor payment history file.
BBC acknowledged that the contract document database revealed that MBE firms received substantially more prime contract dollars in the selected years than the auditor’s vendor payment history file indicated for those years and that their percentage of
*1389
prime contract dollars actually increased after the M/FBE set-aside program was suspended in September, 1989.
See
PS 11-18. BBC says the contract document data is “useful” but asserts its belief that the vendor payment history data “more accurately portrays MBE/FBE utilization as prime contractors in construction.”
Id.
The Court is not convinced that this is the case.
The data in the contract document database was taken directly from construction contract files; it did not rely on approximate classifications of object codes as in the case of the vendor payment history file. The vendor payment history file contains about one hundred and twenty MOCs which relate to goods, services and construction. BBC acknowledged that classification of MOCs was “only approximate” and that MOCs classified as construction might sometimes include nonconstruction purchases. PS 1-13. In creating the contract document database, BBC analyzed 983 contracts and took steps to verify the data. BBC does not give any convincing reasons for preferring the vendor payment data.
The large discrepancies between the amounts of prime contract awards to M/FBEs reflected in the two databases emphasizes the need for complete data and raises questions about the reliability of BBC’s method of extracting construction data from the auditor’s payment history file. Notwithstanding BBC’s contrary opinion, the data from the contract document database would appear to be the more reliable.
If the contract document database is accurate, it leads to conclusions which are significantly different from those which BBC draws from the vendor payment data. The contract document database shows that MBEs received about 4% of the prime contract dollars during the years studied. This is a level of utilization in excess of BBC’s chosen measure of availability, and would negate a hypothesis of discrimination in the award of prime contracts.
See,
PS p. 11-18 and Table II-A-10.
b)
Availability Data
BBC considered three sources of data to determine the number of M/FBE firms “ready, willing and able” to perform construction work for the city. None of these measures of availability purported to measure the number of M/FBEs who were qualified and willing to bid as prime contractors on city construction projects. In order to be selected as a successful bidder on a prime contract, a firm must be deemed responsible and it must be capable of providing a bid bond and a performance bond. The anecdotal evidence collected by the city clearly showed that only a fraction of all construction firms are capable of meeting these requirements. Generally, only large firms with substantial net worth and a significant history of success in the construction business are able to qualify for bonding. BBC never attempted to determine how many firms in the Columbus MSA have such qualifications or how many M/FBE firms in the Columbus MSA are so qualified. There is no basis in the evidence for an inference that qualified M/FBE firms exist in the same proportions as they do in relation to all construction firms in the market. To the contrary, as BBC acknowledged “M/FBEs tend to be smaller and perhaps less likely to have the capabilities to provide goods, services and construction for the City.” PS II-25.
8
The city maintains records of all firms which have submitted bids on prime contracts. This would be a ready source of information regarding the identity of the firms which are qualified to provide contracting services as prime contractors. BBC does not explain why it did not use this data. On prime contracts only the firms which submit bids are “available.”
BBC’s failure to determine the relative availability of M/FBE firms which are qualified to bid as prime contractors is a serious and ultimately fatal omission which renders its conclusions invalid, because, as will be seen, BBC combined prime and subcontract data in conducting its disparity analysis.
i)
Bidder Registration File
One source of availability data BBC considered was the city Bidder Registration File
*1390
(BRF), a list of firms that completed bidder registration forms which were filed with the city’s Purchasing Division. This source of data has several limitations which renders it essentially useless for determining the total number of construction firms which are qualified to provide construction services to the city and the proportion thereof which are M/FBEs. The listing of firms was not limited to the Columbus MSA, and thus does not conform to BBC’s definition of the relevant geographical market. Of the 6,000 goods and services firms contained in the BRF, only 203 are construction firms. This is only a small fraction of the total number of construction firms in the Columbus MSA, as shown by U.S. Census Bureau data. The BRF was not normally used by city departments to solicit construction bids. Thomas Devoe, head of the city’s Division of Public Service, testified that the BRF was only used to solicit bids for contracts under $10,000. The city’s MFBD Division actively recruited minority firms to submit BRF registration forms but did not recruit majority firms. The city’s M/FBE set-aside program was an incentive for M/FBE firms inside and outside the Columbus MSA to submit BRF registration forms. At trial, David Keen of BBC conceded that the number of M/FBEs listed in the BRF was influenced by the previous set-aside program. BBC did not verify the accuracy of the BRF file. Plaintiffs showed that in several instances firms were incorrectly categorized as construction firms.
BBC acknowledged that “[bjecause of these problems ... the Bidder Registration File is a poorer gauge of firms ready, willing and able to provide construction for the City_” PS 11-24. When BBC undertook its disparity analysis by comparing utilization with availability, it chose, in the construction area, to reject the BRF as a source of reliable availability data, stating that “because of the limitations of the Bidder Registration File in measuring availability of construction firms, we do not believe that these disparities represent conclusive evidence of discrimination for construction.” PS 11-30.
The percentage of MBE firms listed in the BRF was 23.7% and the number of FBE firms was 10.9%. The 1987 Census Data indicated that MBE firms comprised only 3% of all construction firms and FBE firms 4.8%. The BRF grossly exaggerates the number of available M/FBEs in proportion to the total number of available construction firms. The BRF is not an appropriate measure of the number or percentage of firms which are qualified to provide construction services to the city.
ii)
Census Data
The second source of data BBC considered in determining availability was 1987 U.S. Bureau of Census data for the Columbus MSA. Census data which identifies firms owned by race and gender is available only for propri-etorships, partnerships and Subchapter S corporations. BBC estimated the number of Subchapter C corporations and assumed they were majority owned. BBC used this census data to estimate the total number of construction firms in the Columbus MSA and the number of M/FBE construction firms. It is apparent, however, that not all construction' firms in the Columbus MSA are qualified, willing and able to bid on city construction contracts. Nevertheless, BBC concluded that this was the best data available to calculate availability of M/FBE construction firms. This data showed that about 3% of all construction firms were MBEs and about 5% were FBEs.
iii)
Telephone Survey
The third source of availability data BBC considered was a telephone survey it conducted of construction firms in the Columbus MSA, using a list it purchased from an unidentified vendor. BBC decided not to use this data as a measure of availability except to note that it supported the reliability of the census data. PS 11-26.
c)
Disparity Analysis
BBC limited its disparity analysis to the eighteen-month period between January, 1990 and June, 1991. BBC said that it limited its disparity analysis to this eighteen-month period because the city’s M/FBE set-aside program was in effect from 1983 to 1989.
*1391
Instead of conducting a separate disparity analysis of prime and subcontract awards, BBC combined them. This is mixing apples and oranges. The qualifications for prime and subcontractors are significantly different, and the decisionmakers and the methods of selection are entirely different. Prime contractors are required by city ordinance to provide bid bonds and performance bonds, subcontractors are not. Surety bonds are available only to those firms which are able to meet stringent financial requirements. Prime contracts are awarded by city officials to the lowest responsive and responsible bidder. Contracts over $10,000 must be approved by city council and the mayor. Subcontracts are privately negotiated between individual prime contractors and subcontractors. The anecdotal evidence collected by the city showed that prime contractors decide with whom they want to work. Bidding may be formal, informal or nonexistent. Subcontracts are often awarded on thé basis of past experience and personal relationships.
BBC’s decision to limit its disparity analysis to combined prime and subcontract data was unwarranted and unjustified. If, as BBC itself acknowledged, fewer M/FBEs are qualified to bid as prime contractors, and therefore their proportionate share of large prime contracts is less than their proportionate share of smaller subcontracts, then combining the prime and subcontracting data would skew the data in the direction of unde-rutilization of M/FBEs.
One of the principal purposes of the EBO Code is to establish annual subcontracting goals for M/FBEs. Such a program must be justified by evidence of discrimination in subcontracting. A statistical analysis which combines prime and subcontract data prevents a separate analysis of M/FBE participation in subcontracting. This methodology defies logic.
In calculating the revenues of M/FBEs for 1990 and 1991 for the purposes of its disparity analysis, BBC excluded revenues from contract modifications.
See,
PS 11-29. The reason it gave for doing so was that many of these were modifications of contracts from prior years. However, revenues from contract modifications were included in calculating the shares of M/FBEs and majority firms in the analysis of the years prior to 1990. Excluding them from the eighteen-month period chosen for the disparity analysis may exaggerate any apparent decline in M/FBEs shares for that period.
Even though census data probably overstates the proportions of available M/FBEs, and even though combining prime and subcontracting dollars would skew the analysis toward a finding of underutilization of M/FBEs, BBC’s disparity analysis for 1990 and the first six months of 1991 failed to demonstrate any disparity for MBEs. BBC’s calculations showed MBE utilization at 7% compared to availability of 3%. For FBEs, BBC’s calculations indicated utilization at the level of 2.5% versus availability of 4.8%.
BBC acknowledged that its methodology revealed that MBE utilization in 1990 and 1991 exceeded availability based upon U.S. census data, but it focused its attention instead on the “rapidly diminishing” utilization concluding that “the findings for construction are difficult to interpret.” PS 11-31.
It is apparent that a disparity analysis of subcontracting for the years the set-aside program was in effect would have shown substantial overutilization of MBEs or reverse discrimination against female and majority subcontractors. In some years MBEs received 30% to 50% of all subcontract dollars. If, as BBC’s data clearly shows, utilization of MBEs as subcontractors was artificially inflated by the previous set-aside program, this would explain why a significant reduction in utilization occurred immediately following the suspension of that program.
The Predicate Study does not provide convincing statistical evidence of discrimination against M/FBEs in city construction.
4.
Predicate Study Supplement
In September of 1993, BBC provided the city with the Predicate Study Supplement, which contains an analysis of data for city construction contracts from June 30, 1991 through May 31,1993.
*1392
In the Predicate Study supplement, BBC reports that MBE utilization as prime contractors declined from 7.7% in 1990 to 5.2% in 1991, to 0.6% in 1992, and then increased to 0.7% in the first five months of 1993. PS Supp. Table 2. There was no significant decline for FBEs. As was the case with the Predicate Study, BBC did not determine the number of M/FBE firms which were qualified to perform construction work for the city as prime contractors. Thus, there is no data on the percentage of available M/FBE prime contractors to compare with the prime contractor utilization percentages. As a result, no conclusions can be drawn about the level of utilization of M/FBEs as prime contractors.
The subcontracting data reported in the Predicate Study Supplement indicates that while utilization of M/FBE subcontractors decreased after the termination of the previous set-aside program their share of subcontracting dollars still substantially exceeded availability as measured by census data.
See,
PS Supp. Table 4. MBE firms received 8.1% of all subcontract dollars in 1991, 6.7% in 1992 and 5.7% in the first five months of 1993.
Id.
The amounts received by FBE firms fluctuated widely from 1.8% in 1991, up to 13.0% in 1992 and down to 0.2% in the first five months of 1993 but for the entire three and one-half year period, their share averaged 9.2%. It will be recalled that in the Predicate Study, BBC concluded that the 1987 census data was the most accurate measure of availability. That data indicated that 3% of available firms were MBEs and 4.8% were FBEs.
After presenting separate data for prime contracts and subcontracts, BBC then combined the data and conducted a disparity analysis based on the combined figures. For the reasons previously stated, the court rejects this methodology.
Combining prime and subcontract awards, BBC calculated MBE utilization rates of 5.2% in 1991, 2.5% in 1992 and 2.4% for the first five months of 1993. PS Supp. Table 5. Although the court believes that combining the data in this fashion skews the results toward a finding of underutilization and is ultimately invalid because BBC never determined the availability of M/FBEs as prime contractors, nevertheless, this methodology shows overutilization of MBEs in 1991 and only a slight degree of underutilization in 1992 and the first five months of 1993,
i.e.,
differences of half a percentage point and six-tenths of a percentage point. This methodology indicates that FBE utilization was .7% in 1991, 4.9% in 1992 and 1.2% in 1993.
The Predicate Study Supplement does not provide convincing statistical evidence of discrimination in the award of city construction contracts.
5.
Predicate Study Second Supplement
In January of 1995, over a year after the city enacted the EBO Code, BBC provided it with a second supplement to the Predicate Study. This report analyzes data relating to the city’s purchase of goods and services and purchases of goods and services in the private sector for firms owned by specific racial and ethnic groups, including blacks, Hispanics, Asians and Native Americans. Thus, the Predicate Study Second Supplement presents availability and utilization data for distinct racial and ethnic groups which were not available in the earlier studies. In this report, BBC analyzes city construction contracting data for the same period covered in the first supplement, to wit: June 30, 1991 through May 31,1993.
At page II-9 of the Second Supplement, BBC states that data regarding the ownership of some firms had changed since its earlier studies. Several firms which were previously reported as majority-owned were reclassified as female-owned and two firms previously reported as minority-owned were reclassified as female-owned. BBC also reported that it had obtained ownership status information for several prime and subcontractors which was not available for the earlier studies.
These adjustments resulted in significant changes in the utilization data for FBEs. Their share of prime contract awards in 1992 were reported in Table 2 of the Predicate Study Supplement as 1.1% of the total, whereas in Exhibit II-8 of the Second Supplement, they are reported as 2.95% of the
*1393
total. The reported share of FBEs of combined prime/subcontraets for 1992 was raised from 4.9%, PS Supp. Table 5, to 6.68%, PS 2d Supp. Ex. II-8. The reported share of FBE subcontracts between January and May, 1993, grew from .2%, PS Supp. Table 4, to 5.11%, PS 2d Supp. Ex. II — 9. Their share of combined prime/subcontraets rose from 1.2%, PS Supp. Table 5, to 2.75%, PS 2d Supp. Ex. II-9. Instead of dropping off to almost nothing in the first five months of 1993, as originally reported by BBC, it appears that FBEs’ share of the subcontracting market remained a robust 5.11%. These corrections to the data in the Second Supplement show just how much small errors in collecting data can impact final utilization and disparity calculations in a data set consisting of relatively small numbers of firms and contracts worth large amounts of dollars. This further underlines the danger in drawing conclusions from incomplete data or from data which covers relatively short periods of time. Indeed, it easts doubt on the propriety of investigating discrimination in construction contracts by resorting to a statistical analysis based on contract dollars.
In the Second Supplement, BBC for the first time reports data on the number of contract awards in addition to the dollar amounts. This new data makes it possible to analyze the proportionate number of contracts awarded to M/FBEs. During the period of July through December, 1991, black-owned firms received 10.39% of the prime contracts, 16.67% of the subcontracts and 13.55% of the combined prime/subcontracts; the share of FBEs was 1.3%, 10.26% and 5.81%, respectively. PS 2d Supp. Ex. II-7. During the year 1992, black-owned firms received 7.48% of the prime contracts, 8.44% of the subcontracts and 8.07% of the combined prime/subcontraets. The share for FBEs was 6.12%, 7.59% and 7.03% respectively. PS 2d Supp.Ex. II — 8. During the period January through May, 1993, black-owned firms received 4.44% of the prime contracts, 6.35% of the subcontracts and 5.56% of the combined prime/subcontraets. The share for FBEs was 4.44%, 4.76% and 4.63% respectively. (PS 2d Supp.Ex. II — 9).
Considering that black-owned firms represent about 2.25% of all available construction firms in the Columbus MSA according to 1987 U.S. Bureau of Census data, and considering that FBE firms represent about 4.8% of all available construction firms in the Columbus MSA, it appears that except for FBE prime contract awards in July through December, 1991, MBEs and FBEs received more than their proportionate share of prime contracts, subcontracts and combined prime/subcontracts, based on the number of contracts awarded.
. BBC’s analysis of combined prime/subcon-traet awards to black-owned firms showed a utilization rate of 9.27% for the first six months of the study period, which then dropped to 1.4% for the remainder of the study period. This apparent sharp decline resulted from the fact that one black-owned firm received a prime contract in the amount of $1,160,000 during the first half of 1991, causing the percentage of prime contracts awarded to black firms to rise sharply to 10.8% of the total. This likewise substantially elevated the percentage of combined prime/subcontracts to a level which appears to be far greater than the share normally enjoyed by black-owned firms, indeed, a level over four times the percentage of their representation in the overall construction market according to census data. For reasons previously stated, the court has rejected BBC’s methodology which combines prime and subcontracting data. This example again shows how just one contract can skew a statistical analysis based on small numbers of contracts which vary substantially in dollar amounts.
The data reported in the Second Supplement revealed that the share of black-owned firms of the total subcontract dollars was 3.11% for the period July through December, 1991, PS 2d Supp. II — 7, 3.79% for 1992, PS 2d Supp. Ex. II-8, and 3.76% for the first five months of 1993, PS 2d Supp. Ex. II-9. This is well in excess of availability based on U.S.. census data.
a)
Availability Data
In the Second Supplement, BBC presents four measures of availability of M/FBE
*1394
firms: 1987 U.S. census data for industry totals; U.S. census data for industry subsec-tors; BBC’s 1992 telephone survey; and the city’s BRF.
i)
U.S. Census Data for Industry Totals
This is the same data and the same methodology as used in the Predicate Study and the First Supplement. It is based on 1987 census data, but the percentages of availability of M/FBE s differ slightly from the percentages used in the Predicate Study and the First Supplement because estimates were made to adjust for underreporting of Hispanic, Asian and Native American-owned firms, and estimates were made to eliminate minority female-owned firms from the FBE totals published by the Bureau of Census. Using this data and this methodology, BBC determined that MBEs represented 3.08% of all construction firms in the Columbus MSA, that black-owned firms represented 2.27% and that FBEs represented 4.49%. This data indicated that there were 230 black-owned construction firms and 455 construction firms owned by white females. PS 2d Supp.Ex. III-l.
ii)
U.S. Census Industry Subsector Data
BBC did not propose to use this measure of availability in either of its prior studies. This data is based upon unpublished U.S. census data from the 1987 Surveys of Minority and Women-Owned Business Enterprises and limited to enterprises identified by four-digit Standard Industrial Classification (SIC) codes for the construction industry. Because corresponding data on all firms was only available for firms having paid employees, it was necessary to use only the M/FBE data for such firms. BBC also noted that the data for all firms was based on the total number of establishments,
9
whereas the M/FBE data was limited to total firms. Finally, it was necessary to resort to other census data to determine the total number of establishments for certain sectors of the construction industry not contained in the surveys of minority- and women-owned business enterprises. Using this data and this methodology, BBC determined that MBE firms represented 4.65% of all construction-related firms in the Columbus MSA, that black-owned firms represented 3.76%, and that firms owned by white women represented 5.03% of all construction related firms. PS 2d Supp.Ex. Ill— 3. This methodology resulted in a significantly higher representation of black-owned firms than the census data for industry totals (2.27%) and the BBC telephone survey (2.5%).
The Court notes that the SIC codes used to generate this data do not correspond with the SIC codes used by BBC to identify construction-related firms in its 1992 telephone survey.
Com/pare
PS 2d Supp.Ex. III-2 with PS 2d Supp.Ex. III-5. A number of codes used in the telephone survey were omitted from the list of codes used to generate the census industry sector data, including 1622 general contractors — bridge, tunnel and elevated highway construction; 1623 general/special trades contractors engaged in water and sewer mains, pipelines and communications and power lines; 1629 general/special trades contractors engaged in other heavy construction not elsewhere classified; 8741-04 construction management; 0782-04 sod and sodding services; and 0782-04 landscape contractors. On the other hand, certain SIC codes which would seem to relate to materials and services were included in the list used to generate the census industry subsector data but were not used in the telephone survey. These include: 4210 trucking and courier service except air; 5030 lumber and other construction materials; 5063 electrical apparatus and equipment, wiring supplies and construction materials; 5070 hardware, plumbing and heating equipment and supplies; and 7353 heavy construction equipment rental/leasing. BBC does not explain these apparent discrepancies.
The court concludes that census industry subsector data, presented for the first time in the Second Supplement, is less reliable than the industry total census data which BBC presented as the preferred measure of avail
*1395
ability in the Predicate Study and the First Supplement. Clearly, the census industry total data is the most comprehensive data, even though it may overestimate the number of firms qualified to provide construction services to the city.
iii)
Telephone Survey
BBC did not use its telephone survey as a measure of availability in the Predicate Study or the Second Supplement. It indicated only that the results of the telephone survey tended to confirm the reliability of the industry total census data. In the Second Supplement, BBC proposes its telephone survey, PS 2d Supp.Ex. Ill — 4, as a measure of availability by selecting only those firms which indicated that they had the ability and capacity to perform one or more of the kinds of construction services the city usually needed. PS 2d Supp. Ill — 15. However, the survey question relating to the ability and capacity to perform building construction or repair projects did not exclude residential construction or repair.
10
Respondents were not asked about their ability to comply with city bonding and insurance requirements. They were not asked whether they had ever performed any work for the city in the past or whether they had ever submitted a bid on a city contract. They were only asked if their firm would be interested in future city work if it were available. The telephone survey has questionable value in determining whether a respondent is qualified to provide construction services to the city. It is of no value in determining whether the respondent is qualified to bid on prime contracts. Nevertheless, using this data and this methodology, BBC determined that MBEs represented 3.53% of all firms in the Columbus MSA that were interested in doing work for the city, that black-owned firms represented 2.45% of such firms and that FBEs represented 11.32% of such firms. PS 2d Supp.Ex. Ill — 6. The percentage of FBEs who responded to the telephone survey is far out of line with the census data and appears to be an aberration. The source of the list BBC used to conduct the survey was not identified or authenticated. The court concludes that the telephone survey is not a rehable source of data for availability of FBEs.
iv)
Bidder Registration File
BBC explained the weaknesses of this potential measure of availability in the Predicate Study. In the Second Supplement, BBC again acknowledged that the BRF may be a poor measure of availability for construction.
See,
PS 2d Supp. III-23. BBC reported that it found that less than one-half of the firms included in its prime/subeontraetor database were listed in the BRF. As the court has already noted, the BRF is not a valid measure of firms which are ready, willing and able to provide construction services to the city, or the proportion thereof which are MTFBEs.
See, supra,
pp. 1389-1390.
b)
BBC’s Conclusions Regarding Appropriate Measures of Availability
At page III-26 of the Second Supplement, BBC states that “all methods of determining availability contain strengths and weaknesses; no one measure emerges as the single strongest approach to examining availability.” This is contrary to the position taken in the Predicate Study, where BBC concluded for good and sufficient reasons that the BRF “is a poorer gauge of firms ready, willing and able to provide construction for the City”, PS 11-24, and that the census data was the best data available to calculate the availability of M/FBEs. In the Predicate Study, BBC said that it chose the census data over its telephone survey data because the census data is “more comprehensive.” PS 11-26. BBC advances no reason for changing its position. BBC’s assertion that all four measures of availability contain strengths and weaknesses and that no one approach emerges as the single strongest approach lacks credibility. The four methods of calculating availability are inconsistent and contradictory, they can
*1396
not all be correct or equally reliable. By failing to make a choice between these four measures of availability, BBC not only contradicted its original position but essentially abandoned its role as an expert in statistical analysis and invited speculation about the existence of discrimination depending on which measure of availability one might choose. The Court can perceive of only one reason why BBC would do this and that is that the measure of availability it originally chose as the most reliable in the Predicate Study did not show underutilization of MBEs in the subsequent study as it anticipated it would.
The court concludes that the census industry subsector data, the BRF and the telephone survey are not appropriate measures of M/FBE availability. While the census total industry data has limitations, it appears to be the best data considered by BBC for use in determining availability of M/FBEs as subcontractors. BBC did not consider any valid method for determining availability of M/FBEs as prime contractors.
c)
Disparity Analysis
In Section IV of the Second Supplement, BBC presents a disparity analysis using the availability and utilization data discussed above. The disparity analysis begins with an analysis of overall M/FBE utilization for all city purchases of goods, services and construction. This analysis includes consideration of the number of purchases, as well as the dollars of purchases. BBC reports that disparities were identified regardless of whether utilization reflected number of purchases or dollars of purchases. However, BBC did not analyze the separate industries in this fashion, nor did it explain its failure to do so. Apparently it considered an analysis of the number of purchases of some significance, and clearly it had the capability of analyzing each market separately. The court has done this analysis for the construction market, using the data on individual contract awards presented in Section II of the Second Supplement.
See, supra,
p. 1393. This revealed that M/FBEs received more than their share of individual prime and subcontract awards when compared with availability as shown by the 1987 census total industry data.
The disparity analysis presented in the Second Supplement to the Predicate Study has no more value than those contained in the Predicate Study and the Predicate Study Supplement. It, too, fails to include a measure of the availability of M/FBEs as prime contractors, and again, BBC combines prime contracting and subcontracting dollars. Nevertheless, keeping in mind that BBC’s flawed methodology is skewed in the direction of underutilization, it is worthwhile to examine the results it produced.
The first data presented in the disparity analysis shows that the share of MBEs of all prime contract dollars as reflected by the vendor payment history file for the entire study period of July, 1991 to December, 1993 was 3%. PS 2d Supp.Ex. IV-3. This is almost exactly equal to the census total industry availability rate of 3.1%. The share of MBEs of combined prime/subeontract dollars for the period July to December, 1991, was 9.8%.
Id.
This is over three times the census total industry rate of availability. Combined prime/subeontract awards for 1992 and the first five months of 1993, were 2.6% and 2.2%.
Id.
While the court does not accept BBC’s methodology of combining prime and subcontract awards, nevertheless, even using this methodology, it is apparent that the rate of supposed underutilization for 1992 is not large. The difference is somewhat larger for the period January through May, 1993, but still less than one percentage point. PS 2d Supp.Ex. IV-7. This should be viewed in light of the fact that the year before BBC’s methodology, which is skewed toward underutilization, revealed instead ov-erutilization on the scale of three times availability. PS 2d Supp.Ex. IV-6.
In the second supplement, BBC did not present the results of a disparity analysis on subcontract dollar awards. Nevertheless, the data presented in Exhibits II — 6, II-7 and II-8 of the Second Supplement show that utilization of M/FBEs on subcontracts remained above availability as measured by census total industry data.
In Ex. IV-4 of the Second Supplement, BBC presents a comparison of FBE utiliza
*1397
tion and availability using prime contract data from the vendor payment history file and combined prime/subcontract data from the contract document database. The share of FBEs of prime contract dollars reflected by the vendor payment history file for the entire study period of July, 1991 to December, 1993, was 1.5%. Their share of combined prime contraeVsubcontract dollars for July through December, 1991, was 1.8%. For the year 1992, their share jumped to 6.7%, and for the first five months of 1993, it fell back to 2.8%. Using the data reported in Exhibits II-7, II-8 and II — 9, the court has calculated the share of FBEs of total combined prime/subcontract awards for the entire period July, 1991 through May, 1993 and finds that the share for FBEs was 5.3%, well in excess of the census total industry availability rate of 4.5%. The court further notes that the data reported in these exhibits indicates that the share of FBEs of subcontracts was 4.35% for July through December, 1991, 13.18% for 1992 and 5.11% for the first five months of 1993, all well in excess of the census total industry measure of availability.
The data presented by BBC in the Second Supplement shows that the percentage of subcontract dollars awarded to black-owned firms during the study period was as follows: July through December, 1991, 3.11%; 1992, 3.79%; and January through May, 1993, 3.76%. PS 2d Supp.Exs. II-7, II-8 and II-9. BBC reported that the number of black-owned firms in the Columbus MSA, as determined by the 1987 census data was 2.25% of all construction firms in the Columbus MSA, and in the Second Supplement this was raised to 2.27%. BBC’s telephone survey revealed that black-owned firms represented 2.5% of all construction firms interested in city work. Based on these measures of availability, there is no statistical evidence of un-derutilization of black-owned firms in city subcontracting.
BBC’s disparity analysis of combined prime/subcontract dollars not surprisingly purported to show underutilization of blaek- and female-owned firms for certain periods of the study. See, Exs. IV-6, IV-7 and IV-8. With respect to the reported disparity for black-owned firms for the year 1992, BBC reported that the difference was not sufficient to reject the hypothesis that chance could explain the difference. Likewise, according to BBC, the reported disparities for black-owned firms and FBEs for the period January through May, 1993 was not sufficiently large to reject the hypothesis that chance could explain the difference — using the census total industry measure of availability. So even BBC’s skewed methodology would not, even if accepted at face value, constitute compelling or convincing statistical evidence of discrimination in the award of city construction contracts.
As the court has repeatedly noted, BBC’s methodology of combining prime and subcontract awards skews the analysis toward a finding of underutilization. Nevertheless, using the data reported in Exhibits II — 7, II-8 and II-9, the court has calculated the share of black-owned firms of total combined prime/subcontract awards for the entire period of the study, July, 1991 through May, 1993 and finds that their share was 2.7%, a rate above the level of availability measured by U.S. Census Bureau total industry data.
The Predicate Study Second Supplement does not contain convincing statistical evidence of discrimination against M/FBEs in the award of city construction contracts. Curiously, in the Second Supplement, BBC did not update the construction utilization data contained in the First Supplement, which was nineteen months old when the Second Supplement was presented to the city.
6.
Decline in Utilization of M/FBES After Termination of Previous Set-Aside Program
Citing a decline in utilization from 10.1% in 1990 to 3.8% in the first six months of 1991, BBC hypothesized in the Predicate Study that “Based upon this trend and anecdotal information from the city, we believe utilization [of MBEs] from mid-1991 through mid-1992 might have declined below 3.0%.” PS 1-33,1-34. BBC recommended that the city improve its efforts to collect utilization data for a future analysis of data from mid-1991 through mid-1992. PS 1-34.
BBC had the opportunity to test its hypothesis that utilization of MBEs may have
*1398
declined below 3.0% after mid-1991 when the city commissioned it to prepare a supplement to the Predicate Study. BBC completed the Predicate Study Supplement in October, 1993. This study analyzed city construction prime and subcontract data for the period July 1, 1991 through May 31, 1993. BBC reported that the utilization of M/FBEs had continued to decline. It reported that “MBE utilization as subcontractors declined from 8.1% of total reported subcontract dollars in 1991 to 5.7% for January through May, 1993.” PSS, p. 5. However, 5.7% is still well in excess of the availability rate of 3.0% as measured by census total industry data. BBC concluded that the utilization of M/FBEs during the period studied in the Predicate Study Supplement was still artificially inflated by the previous set-aside program and by state and federal affirmative action programs, although the influence was declining.
When the city enacted the EBO Code of 1993, it specifically relied on the decline in utilization of M/FBEs since the termination of the previous set-aside program:
Since the city discontinued minority and female business enterprise affirmative action programs in 1989, there has been a significant and steady decline in the participation of African-American and women-owned construction firms in city construction contracts.
EBO Code § 1(D). After the city enacted the EBO Code, BBC provided the city with the Predicate Study Second Supplement. The second supplement was completed in January of 1995 — a year and a half after the end of the period studied in the first supplement. An analysis of construction prime and subcontract data for that period would have supported or refuted BBC’s hypothesis that the utilization of M/FBEs would further decline as the effects of the previous set-aside program continued to diminish. Since the first supplement failed to show underutilization of MBEs in subcontracting three and one-half years after the suspension of the previous set-aside program, data for the next year and a half would obviously be important to the statistical analysis of discrimination in subcontracting. Nevertheless, BBC inexplicably failed to update its statistical analysis of construction data in the Second Supplement.
The court’s deliberations in this case have been prolonged due to the size and complexity of the record, and unfortunately they have also been delayed by several unavoidable interruptions. When it became apparent that another two years of statistical data might be available by the time the court was prepared to render its decision, the court
sua sponte
raised the issue of whether the record should be reopened to admit evidence of M/FBE utilization in city construction after May 31, 1993. (See Order of May 7, 1996). The city’s response was less than enthusiastic. Its counsel asserted that it would take several months to compile the data. The city said that it had the data for goods and services but not construction:
[W]e have that kind of information in the form the Court suggested for goods and services, which is a program that was not challenged in this action and which is on a data base and could be easily and readily produced, but we do not have it for construction prime contracts or subcontracts.
Mr. Keen is present here today pursuant to the Court’s order, but from discussions with him and with Gwen Rogers, who’s the executive director of equal business opportunity office for the city, that what Mr. Keen would need to do is to go out and collect the raw data from the divisions of the city, such as public utilities, public service, and go through those contracts, much as he did in the original instance, and document the dollars, to whom they went, both the prime and the subs on the projects, and then put the data in the form of the data base and present it to the Court.
May 23, 1996 Hrg., Tr. p. 5. The city’s statement strains the court’s credulity. At page 1-34 of the Predicate Study which was delivered to the city in August, 1993, BBC stated, ‘We recommend that the City improve efforts to collect utilization data....” PS 1-34. Even prior to this, Beatty and Roseboro had recommended that the city “collect statistical evidence on subcontracts awarded
and
amounts actually paid to
*1399
M/FBEs by prime contractors....” Beatty Report, p. 4.
In the Predicate Study, BBC reported that the city required prime contractors to complete a subcontractor’s report which was forwarded to the MFBD Division. Two separate city agencies, the MFBD Division and the Legislative Research Office, tabulated the information contained in these reports. PS II-7. These procedures were in effect during the preparation of the Predicate Study which included the period January, 1990 through June, 1991. The Predicate Study Supplement, which was completed in September of 1993 and covered the period July 1, 1991 through May 31, 1993, indicates that the city was still collecting this information at that time. See PSS, p. 1. ■
The EBO Code itself provides that the city will prepare reports of M/FBE utilization on a quarterly basis and prepare annual supplements to the tables contained in the Predicate Study and the Predicate Study Supplement, including Table II-A-11 which contains data on construction subcontract awards. In his trial testimony, council President Kennedy testified, in response to questions from the city’s counsel, that he understood that the collection of this information would be an ongoing process and that it would not stop with the enactment of the EBO legislation.
This court’s order of January 25, 1991, which declared the previous set-aside program unconstitutional, did not restrict the city from obtaining information regarding the race and gender of construction subcontractors. Indeed, to the contrary, the court’s order of January 25, 1991, p. 3, specifically provided: “The city may require such information from contractors after the award of the contract.” In fact, the city did continue to collect this kind of information long after that order. If the city ceased collecting this information after the completion of the first supplement to the Predicate Study, then it did so deliberately, and it should not be heard to argue that race- and gender-based preferences in subcontracting can be justified on the assumption that utilization of M/FBE subcontractors fell below the level of availability after the city stopped collecting the data which would prove or disprove the hypothesis.
The court is left with statistical evidence that covers a period of only three and one-half years. That data shows a decline in' utilization of M/FBEs after the suspension of the previous set-aside program in 1989. It cannot be inferred that this decline is attributable to discrimination. It is more likely that it is attributable to the fact that the previous set-aside program resulted in substantial overutilization of MBEs. Furthermore, although there has been a decline in utilization, the data shows that as of May, 1993, utilization of MBEs in subcontracting has remained well above availability. The suggestion that it has continued to decline to a level of underutilization is not supported by the evidence. The city did not update its statistical proof in the Predicate Study Second Supplement which was completed in January, 1995, nor did it welcome the court’s invitation to update its statistical proof in May of 1996. It would be sheer conjecture to conclude that a statistical analysis of city construction for the period June, 1993 to date would show underutilization of M/FBEs in city construction.
7.
Employment Predicate Study
In August, 1992, the city commissioned BBC to undertake a study for the purpose of determining whether there was evidence of race or gender discrimination in employment in the goods, services and construction industries in the Columbus MSA or among the city’s suppliers and contractors which would justify legislation containing race- and gender-based remedies designed to increase the employment of minorities and women in the work force of contractors and vendors doing business with the city. Thereafter, BBC presented the city with a report setting forth the results of its research and analysis of statistical and anecdotal evidence of discrimination in employment. This report is entitled “Employment Predicate Study” (“Employment Study” or “ES”). The report concludes with recommendations for race-conscious and race-neutral measures affecting employment by city contractors, but apparently to date the city has not acted on
*14

[Text truncated at 120,000 characters. The full text is on the page linked above.]

---

Source: Frix Law Library, https://www.frixlaw.com/law-library/cases/1487035. Public record. Not legal advice.