Although economists and sociologists continue to dispute the role of discrimination law in contributing to blacks' economic progress in the 1960s and 1970s, [FN1] there is an emerging consensus that the "success story" of that period is unlikely to be replicated in the next decade. [FN2] The circumstances that enabled the law to be effective during that period are unlikely to recur. Several factors are of particular importance. First, the overt use of discriminatory criteria--often based on Jim Crow laws--has been dismantled and the less overt, "neutral" barriers to minorities' employment-- the battery of apparently discriminatory "objective" performance tests--have been removed either as invalid or as too costly to validate in court. Further progress in eliminating discrimination must aim at forms of discrimination that are more subtle, and therefore more difficult to eradicate. [FN3] Second, the manufacturing sector, which made great progress through quotas and training programs, is shrinking and so offers diminished opportunities to young minority workers first entering the labor market. [FN4] Third, it appears unlikely that the success of discrimination law for relatively routine jobs in the manufacturing sector can be extended to high- level jobs--those professional and technical jobs that form the desirable core of the new information economy. [FN5] In particular, the legal system cannot effectively second-guess the validity of managers' personnel decisions in jobs where job performance is evaluated on subjective performance criteria. [FN6]

 
We contend here that the central challenge to further success of the anti-discrimination norm is the difficulty of applying its legal conceptions to jobs where performance is difficult to monitor. Our central claim is that discrimination is likely to be self-perpetuating (in part because of actions and choices by employees) in these jobs, and that current law, with its exclusive focus on intentional discrimination by the employer, is likely to be ineffective in eradicating this self-perpetuating discrimination.

 
The core of the argument is an analysis of how discrimination affects the job application process. In the legal literature, the most influential economic analysis of the anti-discrimination norm generally has depended upon three central analytic features: (1) a simple market-clearing model for labor markets, in which wage equals marginal product, and wages fall until the market clears; (2) a homogeneous view of the firm, in which information about the qualities and efforts of workers is perfect and in which hiring and firing is costless; and (3) a conception of discrimination as simply animus-based. [FN7] These assumptions, although permitting legal analysts to reach incisive conclusions on the basis of relatively simple and rigorous economic models, [FN8] are inconsistent not only with ordinary observations of the workplace and of labor markets, but also with the generally received view among economists of how labor markets work. Conventional economists now see that this simple model is seriously at odds with the operation of many sectors of the labor market. [FN9]

 
For these sectors, labor economists have developed a variety of "efficiency-wage" and "tournament" models. [FN10] These models share two essential features. First, firms find it profitable to set wages higher than the market-clearing wage, creating a pool of unemployed workers from which they can draw. [FN11] Second, because it is costly to monitor or fire workers once they are hired, firms must adopt complex hierarchical structures that provide incentives for enhanced worker effort. [FN12] These two features of a sophisticated microanalytic view of the firm suggest a fundamentally different view of discrimination law from those based on market-clearing models and homogenous labor conditions across firms.

 
As a methodological matter, the key claim of this Article is that the organizational structure of firms, and the impact of this structure on the labor markets in which these firms hire, is an important feature in analyzing the causes of employment discrimination and the effects of anti-discrimination law. [FN13] Instead of resting on broad generalities or abstractions about features of the employment markets and the policies of employers, discrimination analysis should extend to a more refined understanding of work structures, such as definitions of job qualifications, patterns of promotion, degrees of inter-firm mobility by workers, and compensation policies. These structures powerfully affect the firms' propensity to discriminate and workers' responses to substantial sectors of the workplace. This Article focuses on high-skill or knowledge intensive jobs in which the discretionary, skill-based component of work creates substantial difficulty in monitoring individual workers' performance. The analysis we develop applies, for example, to law firms, investment and commercial banks, brokerage houses, management and engineering consulting firms, accounting firms, universities, and governmental and large corporate bureaucracies. [FN14]

Our analysis calls for a rethinking of current discrimination law and policy. Current anti-discrimination doctrine is fundamentally ill-suited to rooting out the types of unequal opportunity that our analysis identifies. Attempts by certain judges to remedy the problem by tweaking the pleading requirements and burdens of proof under the existing case law will not eradicate discrimination in the long run. We suggest that the best approach is to restructure the law to induce employers to take a proactive role in eliminating inequality. This would involve restructuring compensation, job assignments, and career pathways to provide equal opportunities for professional achievement to all employees.

                      FNa1. Professor of Law, Harvard Law School.

The Third Circuit Court of Appeals, sitting en banc, decided Taxman v. Board of Education of the Township of Piscataway, [FN1] in August 1996. Eight judges agreed that the Board of Education of Piscataway Township, New Jersey violated Title VII of the Civil Rights Act [FN2] by using race, in accordance with its affirmative action policy, to break a tie between two teachers in the Business Department at Piscataway High School when determining which teacher to lay off. [FN3] A strong dissent by Chief Judge Sloviter was joined by two other Court of Appeals judges. [FN4] The majority decision is remarkable in its breadth, concluding that Title VII permits race-based decision-making only for the narrow purpose of remedying the present effects of past discrimination. [FN5] This reading of Title VII strongly implies that Title VII may limit a public institution's race-based decision-making beyond the limitations imposed by the Equal Protection Clause of the Fourteenth Amendment. [FN6] Not one Supreme Court Justice has ever articulated this position before. In fact, the Court in Johnson v. Transportation Agency [FN7] concluded that Title VII gave greater latitude to public employers to use race-based voluntary affirmative action measures than did the Equal Protection Clause. [FN8] In Taxman, the Board of Education petitioned for certiorari, which the Supreme Court granted on June 26, 1997.

A number of curious happenings demonstrate the political importance of the case. These include the amount of press coverage accorded to the case, [FN9] the remarkable settlement that avoided Supreme Court review [FN10] and the flip-flopping of the United States government's position on the case. [FN11]

The case, which was scheduled for oral argument in January, 1998, settled on November 21, 1997. [FN12] Remarkably, civil rights organizations, in particular, the Black Leadership Forum, concerned that the case could make bad precedent and virtually abolish affirmative action programs, contributed significant funds to settle the case. [FN13]

The role of the federal government in this case was also unique. Originally the United States Equal Employment Opportunity Commission (EEOC) filed the suit against the Board of Education, alleging race discrimination in violation of Title VII, but Sharon Taxman intervened. Both the EEOC and Taxman filed motions for partial summary judgment, which the district court granted. On the issue of damages, the jury awarded over $144,000 to Sharon Taxman.

The Board of Education appealed to the Third Circuit Court of Appeals. The United States, shifting sides, moved to file a brief supporting the position of the Board of Education, stating that it could no longer support the judgment of the trial court. The Court of Appeals permitted the United States to withdraw as a party, but denied the government's motion to file an amicus brief on appellant's behalf. [FN14]

The Third Circuit affirmed the lower court's judgment. The Board of Education petitioned for a writ of certiorari; the United States Justice Department filed a brief in opposition to the petition for certiorari. On June 27, 1997 the United States Supreme Court granted the writ of certiorari. [FN15] The Solicitor General of the United States filed a Brief for the United States as Amicus Curiae in which the government argued for affirmance of the Court of Appeals holding, but took issue with the court's reasoning that all non-remedial, race-conscious employment decisions are illegal under Title VII. [FN16]

At the announcement of the settlement I felt contradictory emotions: relief and disappointment. I was relieved that there would be no opportunity for the Supreme Court to abolish affirmative action. I was doubly disappointed, however, because I believed that Taxman presented an opportunity for the Supreme Court to eschew the broad simplistic approach to affirmative action that the Fourth, [FN17] Fifth [FN18] and now, the Third, Circuit Courts of Appeals had taken. If given the opportunity to decide an affirmative action case, the Supreme Court, I believed, would not deal a mortal blow to affirmative action, but would arrive at a middle ground that opponents and proponents of affirmative action could accept.

It was with this mindset that I approached this project. My role as a participant in this project was twofold. First, I was to assume that Taxman had not settled and that the Court had overturned the Third Circuit's decision. Based on this assumption, I was to write the Court's opinion, overturning the Third Circuit Court of Appeals decision and any concurring or dissenting opinions. While writing the opinions, I tried to anticipate how the Justices would decide the case without interjecting my own views into the opinions. I have reviewed the record and briefs of the parties and amici and the affirmative action cases decided by the Supreme Court under Title VII and the Equal Protection Clause, parsing each Justice's view toward affirmative action. I have tried to be faithful to each Justice's view toward affirmative action, procedural issues, employment discrimination, and education law.

My second contribution to the project was to write this commentary, explaining the choices I faced. This assignment gives me leeway to examine how this case looks through the eyes of a Supreme Court Justice and to discuss my own views concerning its substance and procedure. Besides piquing my interest as a legal scholar, this task also raised pedagogical issues about teaching future lawyers. For a short time, I stepped into the shoes of a Supreme Court Justice to see what strategies were successful and unsuccessful. I will try to share with my readers those perspectives as I go along.

A number of fascinating issues arose. They included questions concerning the facts of the case, the respective burdens of proof, summary judgment law, the proper interpretation of Title VII, the interplay between Title VII and the Equal Protection Clause, the various meanings of 'remedial' purpose, whether a non-remedial purpose was sufficient for race-based decision-making under Title VII, whether employment discrimination law differs in an educational setting, whether a layoff rather than a hiring decision would make a difference affecting the outcome, the changing political focus of the federal courts and institutional competence.

Although any one of these matters would provide substance for a separate law review article, I will attempt briefly to address these issues and to raise questions that I hope others will develop.

January 1, 1998
William Jefferson Clinton
The President of the United States
Office of the President
1600 Pennsylvania Avenue, N.W.
Washington, D.C. 20500

Re: OPEN LETTER TO THE PRESIDENT ON RACE AND AFFIRMATIVE ACTION

Dear President Clinton:

As a professor, former U.S. Equal Employment Opportunity Commission (EEOC) Commissioner and concerned citizen, I write to you about race in America. This letter is to contribute to the national race initiative discussion, which you have so boldly initiated, with the hope that my comments will be useful to you and the Advisory Board on the President's Initiative on Race. While my letter primarily focuses on the moral imperative of affirmative action, it will also cover several race-related issues.

As you know, the affirmative action process is one of the most important employment initiatives of this century. Affirmative action "has not outlived its usefulness" [FN1] because racism still impacts our society. For example, a poll by the National Opinion Research Center at the University of Chicago reveals that 53.2% of most Americans believe blacks are less intelligent, 62.2% believe they are less hardworking, and 77.7% believe they have a greater preference for living on welfare than whites. [FN2] In an attempt to minimize the effect of such prejudices, society has created a structure of laws, ethics, and educational programs.

Untold numbers of minorities and whites have benefited from affirmative action in America. It was designed to eliminate de jure and de facto discrimination in the workforce, education and government contract allocation. Because racism, sexism and classism are pervasive in America, certain groups have a dire need for affirmative action. For example, local and national governments have used set-asides and contractual inducement policies to enhance opportunities for minorities shut-out of the process, and to reward companies in the business community when they agree to further the opportunity policies of government. When employers use affirmative action, they do so either voluntarily or when required by the courts.

I. UNDERSTANDING THE CONCEPT OF AFFIRMATIVE ACTION

Affirmative action seeks to achieve at least two objectives: (1) the elimination of barriers that have kept minorities and women out of the economic mainstream, and (2) the initiation of positive measures that ensure a truly equal opportunity for those previously excluded. Both opponents and proponents of affirmative action have a common objective: to get beyond racism. [FN3] They differ on "whether affirmative action is an appropriate means to reach the desired end." [FN4] Affirmative action is one such vehicle. I submit that affirmative action remains good public policy as a remedy to correct the effects of race-based discrimination by government and the private sector. [FN5]

Affirmative action has been the subject of much public debate and confusion in recent years. [FN6] It encompasses a variety of activities, some of which have extensive public support, and others which have been sharply criticized. Affirmative action can be defined as plans, policies or legislative enactments appropriate to overcome the barriers to equal employment opportunity and economic interests in the public and private sectors. Frequently referred to as "a plan of action," affirmative action is no more than a systematic organizational effort to reach certain private or public management objectives based on sound organizational analysis and problem identification. [FN7]

Many experts have clarified the meaning and purpose of affirmative action. For example, former Chairman of the United States Commission on Civil Rights, Republican Arthur S. Flemming was asked: "What do you mean by affirmative action, and why do you feel that it is important?" [FN8] His response suggested that in the public sector affirmative action has both functional and policy meanings. The functional application of affirmative action usually consists of curing a wrong after a court or administrative body has found or adjudicated liability. The policy meaning of affirmative action is derived from a determination by a public body that inequity exists in public programs that the public body chooses to cure as a condition for private participation in such programs. Another kind of affirmative action program is that administered by private companies. These programs help avoid litigation by allowing companies to remedy their own misconduct.

Introduction ....................................................... 1402

I.    Enforcement of the Fair Housing Act ................................ 1405

        A. The Government's Efforts ...................................... 1405
          1. The Complaint Filing Process ................................ 1406
          2. The Administrative Process .................................. 1411
          3. Proceeding to Judgment: Choice of Forum ..................... 1415
        B. Government Efforts Compared to the Private Bar ................ 1416
        C. Pattern and Practice Cases .................................... 1422
          1. Home Mortgage Discrimination ................................ 1423
          2. Testing and Disparate Impact Cases .......................... 1425

II.   Employment Discrimination: The Case of the EEOC .................... 1427

        A. Do Politics Matter? ........................................... 1430
        B. Private Attorneys and the 1991 Act ............................ 1435

III.  Can the Government Effectively Enforce Civil Rights? ............... 1438

        A. Explaining the Government's Enforcement Policies .............. 1439
          1. Behavioral Incentives for Government Attorneys .............. 1442
          2. Theory in Action: The Justice Department's Employment
        Litigation ....................................................... 1447
          3. Other Bureaucratic Constraints: The Government as a
        Defendant ........................................................ 1450
        B. Private Attorneys and Civil Rights Cases ...................... 1452
        C. The Role of the Government in Enforcing Civil Rights .......... 1456

      Conclusion ......................................................... 1458

In this Article, Professor Michael Selmi contends that one important reason civil rights legislation has produced less change than originally expected is that most of the legislation entrusted principal enforcement to the federal government, and that enforcement has been seriously deficient over time. Through a detailed empirical analysis, this Article compares the efforts of public and private attorneys in enforcing fair housing and employment discrimination laws, and demonstrates that the government generally brings far fewer cases and receives substantially less relief than private attorneys. In both housing and employment, the government has concentrated its efforts on individual cases, focusing primarily on family status housing cases and age discrimination employment cases. The discrepancies between the two enforcement groups arise, Professor Selmi argues, from bureaucratic pressures that prod government attorneys to bring easy and noncontroversial cases as a means of avoiding the conflict that so readily accompanies civil rights enforcement.

When the historic Civil Rights Acts were passed in the 1960s, there was great optimism and hope that the legislation would play a critical role in reducing inequality and eradicating the vestiges of America's segregated past. Thirty years later, we know that the optimism surrounding the legislation far exceeded our abilities, and today there seems to be a wide, and growing, consensus within the American legal establishment that law is an inappropriate or, at best, a limited means to promote social change. In his provocative work, Gerald Rosenberg suggests that courts are incapable of producing the societal support necessary for broad social change, and that, rather than garnering support, courts are more likely to promote a backlash among the public that will ultimately impede social progress. [FN1] Other commentators from both the left and the right have recently adopted a similar theme, [FN2] and it seems that virtually no one today contends that law can play a leading role in fomenting social change. [FN3]

The gains prompted by civil rights legislation are admittedly disappointing by almost any measure. That said, I seek to challenge the prevailing thesis that law is incapable of promoting broad social change. In this Article, I suggest that one of the primary, though by no means only, reasons legal reforms have often failed to produce more social change is the way in which civil rights statutes have been enforced. [FN4] Since their inception, primary enforcement of civil rights laws has been entrusted to the federal government and has been largely handled through administrative procedures. This is especially true with respect to the two areas this Article addresses: housing and employment, both of which rely extensively on a federal administrative process to resolve the vast majority of claims that are filed. Given the requirement that a civil rights litigant must depend on the government for relief, the government has been perceived as the primary enforcement vehicle for civil rights. And yet, for a variety of reasons, the government has failed to play a strong role as an enforcement agency. As has been well-documented, this has generally been true of Republican administrations but, as I establish in this Article, it has also been true of Democratic administrations, when the government has repeatedly failed in its role as a vigorous advocate for those seeking to be free from discrimination. In contrast, private attorneys have been responsible for the vast majority of important civil rights cases and, accordingly, have been principally responsible for whatever social change has resulted from legal challenges. [FN5]

Despite the greater potential of the private bar to bring about social change, civil rights legislation was initially structured in a manner that was certain to limit its attractiveness to private attorneys; specifically, damages were unavailable to private plaintiffs in employment cases, [FN6] and only limited damages were available for fair housing violations. [FN7] Recently, damages have become available for both employment and housing discrimination, [FN8] and, as I document below, those damage provisions demonstrate that private attorneys are willing to pursue civil rights cases as long as they are sufficiently lucrative to warrant their investment. Not only did the absence of damages suppress the supply of attorneys willing to take cases, but also the lack of significant financial penalties likely limited the legislation's deterrent effect on employers and housing providers, both of whom had little financial incentive to change their practices. [FN9]

This Article contains a comparative empirical analysis of the effects of private versus public enforcement of civil rights, concentrating on the areas of housing and employment discrimination. [FN10] Reviewing data from a variety of sources, I demonstrate that, on average, the government seeks and obtains less monetary relief for plaintiffs than does the private bar and fails to address cutting edge issues, choosing instead to concentrate its efforts on small, routine cases. Indeed, one reason for the government's lack of success is that its enforcement is almost entirely driven by the complaints that are filed, and the government has adopted no clear enforcement priorities for these complaints other than to process and dismiss them. This results in a rather haphazard enforcement agenda in which there is little evident effort to employ effectively the government's resources. As a result, the government's enforcement strategies seemingly defy logic: The federal government, with its broad resources, pursues small, politically inoffensive, and easy cases, leaving the private bar to tackle the difficult and important discrimination claims.

In the final part of the Article, I explain why it is that the government's efforts have failed, and why those efforts may have limited the efficacy of the Civil Rights Acts. These reasons include the political concerns that have driven the government's enforcement agenda across various administrations, as well as the composition of the government attorneys charged with enforcing civil rights--many of whom are young and inexperienced but nevertheless afforded substantial responsibility for determining the government's agenda. Under such circumstances, these attorneys act cautiously in their case selection because their primary concern is often future employment opportunities that are best enhanced by winning cases or by obtaining trial experience. Moreover, given the political realities of enforcing civil rights laws, the attorneys who remain through different administrations are likely not to have a passion for civil rights, which again limits their enforcement zeal. As a result, I suggest that the government is inherently a weak enforcer of civil rights, and that it may be time for it to cede its role as a primary enforcement agency.

In Part I of this Article, I discuss the enforcement of the Fair Housing Act, including the relative success and accomplishments of the federal government and the private bar. Next, in Part II, I explore the enforcement of employment discrimination laws, focusing on a comparison between the Equal Employment Opportunity Commission (EEOC) and the private bar, especially under the Clinton Administration. Finally, in Part III, I illustrate why the government's efforts have been of such limited success and further suggest that the government's role must be restructured to focus on class action rather than individual cases of discrimination.

 
Table of Contents

I. Introduction ....................................................... 512

A. Franchising Growth and Regulation ............................. 512
B. The Franchisor-Franchisee Relationship and Statutory Law ...... 518
C. Discrimination and Affirmative Action Issues .................. 520

II. Discrimination Claims .............................................. 521

A. Theories: Implied Covenants and 'Good Cause' .................. 521
B. Representative Cases .......................................... 524
C. Antidiscrimination Statutes ................................... 527

III. Racial Discrimination: Section 1981 Claims ......................... 528

A. Minority Applicants or Franchisees ............................ 528
B. White Applicants or Franchisees ............................... 537

IV. Franchisee Selection ............................................... 538

A. Suspicious Overall Statistics, but Little Proof for Particular
Complaints ....................................................... 538
B. A Rare Franchisee Success: The Sud Case ....................... 548
C. Failure to Provide Adequate Information About Minority
Business Failure Rates: The Sperau Case .......................... 549
D. Brown v. American Honda Motor Co.: Representative of the
Dismal Record for Franchisee Selection Discrimination Claims ..... 552

V. Set-asides, Affirmative Action, and International Law .............. 554

VI. Franchise Site Selection and Market Price Determination ............ 559

A. Redlining in General .......................................... 559
B. Excluding White Franchisees from Minority Neighborhoods ....... 561
C. Redlining of Minority Franchisees ............................. 563
    1. The Problem ................................................ 563
    2. Redlining Charges Against Prominent Franchisors ............ 565
    3. Summation .................................................. 571
D. Franchise Market Price Determination .......................... 572
E. The 'Greenlining' Alternative to Redlining .................... 574

VII. Reactions by Both the Public and Private Sectors to Discrimination
Allegations ...................................................... 575

A. Franchisor-initiated Affirmative Action Programs .............. 575
    1. Responses to Discrimination Claims and Outside Pressure .... 575
    2. Affirmative Action Programs in the Fast-food Industry ...... 584
    3. For Franchise Opportunities, Insufficient Efforts and
    Results? The Burger King Example ................................. 589
    4. The 'Nothing Special' Approach to Minority Recruitment ..... 592
    5. Is a Small Franchisor Better for Minorities? ............... 594
B. The Future of Franchisor-initiated Affirmative Action ......... 595

VIII. Proposals .......................................................... 597

A. Required Disclosures to Prospective Franchisees ............... 597
B. A Private System to Counter the Redlining of Franchises and
the Poor Servicing of Redlined Communities ....................... 599
    1. 'High Risk' Franchises ................................... 599
    2. Special Assistance Based on the Franchisee's Economic
    Circumstances .................................................... 602

IX. Conclusion ......................................................... 610

A. Franchising Growth and Regulation

Franchising is, among other things, a system of marketing and distribution whereby a small independent businessperson (the franchisee) is granted the right to market the goods and services of another (the franchisor) in accordance with established standards and practices. [FN1] In franchising's ideal state, 'the franchisor obtains new sources of expansion capital, new

distribution markets, and self-motivated vendors of its products, while the franchisee acquires the products, expertise, stability, and marketing savvy usually reserved only for larger enterprises.' [FN2] By owning a franchise, as opposed to starting a small business, franchisees may substantially reduce the risk incurred by building an enterprise from the ground up, while gaining
significant experience in their field through the assistance provided to them by the franchisor. [FN3] In addition, franchisees usually gain better name and product recognition. [FN4]

Some commentators have called franchising "the most successful marketing concept ever created." [FN5] If sheer growth indicates success, such a statement may not be hyperbole. Franchising has taken an increasing share of domestic and international business. Over half a million franchises operate in the United States, and they now account for about a trillion dollars in annual retail sales, with franchised outlets responsible for well over a third of total U.S. retail sales. [FN6] The number of franchises and business format franchisors, as well as the percentage of total retail sales, have grown rapidly, far outpacing the economy as a whole in the last few decades. [FN7] In addition to this continued increase in market share, the number of persons working for franchised outlets has skyrocketed from an estimated three and a half million in 1975, to about seven million in the early 1990s, to a predicted figure of ten million by the year 2000. [FN8]

Franchising's "success" and resulting rapid growth have caused problems. Franchise business developments continue to outpace any attempts at developing a well-structured, orderly approach to the legal issues associated with franchising. Court decisions, administrative regulations, and legislation have failed to provide even the most basic element for nationwide legal standards: a uniformly accepted definition of franchising. [FN9]

Most states consider a franchise to exist whenever a franchisee, in return for paying a franchise fee, is granted the right to sell goods or services under a marketing plan prescribed by the franchisor. [FN10] The marketing plan must be substantially related to the franchisor's trademark, service mark, or other commercial symbol. [FN11] A minority of states apply a slightly variant definition, substituting for the marketing plan requirement a "community of interest" between the franchisor and the franchisee in the marketing of goods or services. [FN12]

While there is a federal rule requiring information disclosures, [FN13] the only federal substantive laws that presently govern the franchisor-franchisee relationship directly are those regulating gasoline service stations [FN14] and automobile dealerships. [FN15] Many federal franchise relationship bills have died without even a committee vote, [FN16] and the Federal Trade Commission ("FTC") has never modified the information disclosure rule it promulgated in 1978. [FN17] However, many states do comprehensively regulate franchises, often by codifying a "good cause" standard for the franchisor's dealings with its franchisees. [FN18] This approach requires franchisors to make an affirmative showing of "good cause" before terminating or otherwise adversely affecting a franchise. It constitutes a pro-franchisee step beyond simply incorporating into the common law of contracts the implied covenant of good faith and fair dealing. Even if there is an express contractual provision about termination--one that trumps any implied covenant [FN19]--a court looking for "good cause" must examine whether the franchisor had sufficient cause to terminate a franchise. [FN20]

B. The Franchisor-Franchisee Relationship and Statutory Law

Despite franchising's general acclaim, the franchise relationship often has become adversarial. [FN21] Indeed, while the word, "franchise," is derived from the Old French word, franchir, meaning to "free from servitude," [FN22] some critics have concluded that franchising is a modern form of long-term indentured servitude. [FN23] Because franchisors usually have the upper hand, and in recognition of the typical franchisor's superior bargaining power, some statutes and case law seek to intervene in the relationship and protect the franchisee by imposing upon franchisors an obligation of good faith, including a duty not to terminate franchises except for "good cause." [FN24]

A number of states limit the application of a "good cause" standard to terminations, [FN25] while others apply it to both terminations and refusals to renew. [FN26] Restricting termination to "good cause" is generally designed to prevent franchisors from engaging in "opportunistic behavior prior to the end of the contract term." [FN27] Although the standards underlying "good cause" vary among the states, [FN28] definitions generally cite "good cause" as the failure of a franchisee to comply with any applicable law or with any lawful provision of the franchise agreement, after being given the opportunity to cure that defect. [FN29] The cure period may be as short as five or ten days, [FN30] or as long as two months or more, [FN31] with it often in the middle, at thirty days. [FN32] Franchisors object to "good cause" requirements as impediments on their right to contract freely, arguing that trademark licenses in particular must be guarded through strict quality control. [FN33]

C. Discrimination and Affirmative Action Issues

This Article reviews case law and statutes concerning franchisors' alleged discrimination against franchisees. It considers both franchise terminations and instances in which a potential franchisee was not chosen. The Article analyzes statistics concerning the percentages of minority-owned franchises, and it discusses the use of set-asides and the affirmative action framework in other nations, particularly Canada.

This Article then examines franchisor-orchestrated affirmative action programs designed to increase the number of minority-owned franchises. It proposes private systems to counter the redlining of franchises and the poor servicing of redlined communities. These franchisor-implemented systems would better assist new and existing "high risk" franchises: ones located in markets or communities where businesses face higher costs and lower rates of return, often due to such factors as the low per capita income of potential customers and high crime rates. Under the proposal presented here, extraordinary disclosures--beyond those required by the federal rules [FN34]--would be made to prospective "high risk" franchisees. Other assistance could be furnished to economically disadvantaged, prospective, or existing franchisees, regardless of their race, gender, or ethnicity, and despite the fact that they may not operate in a high risk location/market.

A hypothetical employer seeks to fill several positions. [FN1] Hoping to make its selection from the broadest possible pool of applicants, the employer advertises in the classified section of two newspapers: the New York Times and the Amsterdam News, a newspaper with a primarily black readership. In both advertisements, the employer identifies itself as an "equal opportunity employer" and states that "women and minorities" are encouraged to apply.

Another hypothetical employer plans a recruitment trip to various colleges and universities and includes Howard and Spelman, traditionally black colleges, in its recruitment efforts. At all of the institutions at which it recruits, the employer informs potential applicants that it seeks to enhance diversity in its ranks, and that it is taking affirmative steps to make that aspiration a reality.

A third employer adopts strong reporting procedures requiring the organization to monitor the race and gender of those it recruits, those it hires, and those it promotes. That information is forwarded to top management and assists top management in formulating additional recruiting efforts to attract and retain minorities and women at the organization.

None of these employers seeks to fill a numerical quota of minority or female employees. [FN2] Our hypothetical employers do not set-aside particular positions for minorities or women or provide any kind of preference to them in the hiring process. The same standards apply to all candidates for hiring, retention, and promotion, regardless of race or gender. These employers reach out through race, or gender-conscious means to certain groups to include them in the applicant pool, recognizing that the pool may have been narrower in the past because of a variety of factors. These employers are hesitant to rely solely on word-of-mouth referrals and other traditional means of recruitment because they recognize that typical patterns of social interaction and residential segregation may tend to yield a homogenous applicant pool. [FN3] Have these employers presumptively violated the United States Constitution?

Recently, two United States Courts of Appeals handed down fundamentally inconsistent decisions in answer to this question. Both decisions addressed the types of outreach efforts described above, which I call "race-conscious, non-preferential" or "soft" affirmative action. [FN4] Raso v. Lago and Lutheran Church-Missouri Synod v. FCC have been controversial, both in legal circles and in the media. Arguing that they were unfairly disqualified in a tenant selection process for subsidized housing units, the plaintiffs in Raso vowed to "appeal their case all the way to the Supreme Court." [FN5] In July 1998, an influential group of broadcasters and media companies decided to comply voluntarily with the FCC's equal opportunity regulations challenged in Lutheran Church, notwithstanding the D.C. Circuit's ruling invalidating those regulations as unconstitutional; civil rights leaders called the decision "probably the most extreme ruling of any American court in a civil rights case this decade." [FN6]

Why the flurry of controversy? As the Supreme Court has effectively dismantled preferential forms of affirmative action in recent years, affirmative outreach represents the last wave of affirmative action that has remained largely unchallenged in the courts and in legal scholarship, even by vehement foes of affirmative action. [FN7] These forms of affirmative action do not prefer one racial group or gender over another in a competition for scarce benefits, [FN8] set-aside a particular benefit based on race or gender, [FN9] or mandate quotas that must be filled by members of the preferred racial group or gender. [FN10]

Rather, these forms of affirmative action are explicitly non-preferential and seek to even the playing field by promoting the free flow of information to ensure equal opportunity for minority groups and women. [FN11] To this end, these efforts attempt to increase the pool of qualified minorities and women by informing them of a benefit that had been foreclosed to them in the past because of explicitly biased decision-making in areas such as employment or housing. [FN12] Until recently, these forms of affirmative action were thought to be relatively non-controversial [FN13] and necessary for full inclusion of women and minorities in all facets of American life. [FN14]

Recent arguments, however, equate non-preferential recruitment and retention programs with preferential forms of affirmative action, such as setting aside seats in a law school class for minority students who meet different admissions standards than white students, or laying off white teachers before black teachers regardless of seniority. [FN15] If these arguments prevail, the last wave of permissible affirmative action will be held presumptively unconstitutional and likely dismantled. This result would render the ideal of the even playing field an unattainable illusion. This is not mere hyperbole: equal access to information and opportunity is imperative for minorities to compete for the same opportunities as whites. [FN16] By engaging in an argument about whether affirmative outreach to minorities is constitutional, we beg the question of whether we as a society have abandoned the ideal of the even playing field altogether. If we have, we must grapple with the disastrous consequences not only for minorities seeking access to opportunities, but also for our entire society and its ability to survive and compete in the twenty- first century. [FN17]

We must recognize that past discrimination has had an effect on minority access to opportunity structures. While the Supreme Court has indicated that "there can be no such thing as either a creditor or a debtor race," [FN18] it would be naive to assert that race is not a defining factor of life in the United States. In the context of more preferential forms of affirmative action, recent Supreme Court jurisprudence has suggested that the government must be "colorblind" in its dealings with its citizens. [FN19] Race-conscious, non-preferential affirmative action allows decision-makers to strive for colorblindness by recognizing that we must first look at race to get beyond it. [FN20] These programs are unique in that they allow decision-makers to acknowledge prior race-based decision-making, broaden their applicant pool accordingly, and select their candidates of choice without burdening any other group.

The fact that these programs have come under recent attack highlights a troubling undercurrent in the law and our society: in dismantling preferential forms of affirmative action, the courts have told us that minorities cannot be given an advantage over whites. Recent challenges to non-preferential forms of affirmative action suggest a more disturbing notion: that whites do not

want to compete on an even playing field with minorities, and courts believe that the ideal of an even playing field is objectionable in and of itself. The Lutheran Church court noted that the term "diversity" has been asked to bear a great burden in late twentieth-century America. [FN21] The court's decision implies that, in fact, whites have been asked to bear the burden of diversity, and that this burden must be lifted from their shoulders.

Critical race theorists have argued that colorblindness does not create equality between whites and blacks, but rather allows the government to ignore existing power structures and inherited privilege, which operate in favor of whites. [FN22] The application of the colorblindness notion to non-preferential affirmative action has the same effect: to reify existing power relations, and, ironically, to undercut vigorous competition for sought-after benefits. It is critical to recognize that current jurisprudence regarding preferential affirmative action does not mandate the dismantling of non-preferential affirmative action simply because it is race- conscious. It is also imperative that the judiciary distinguish between these two types of affirmative action, so that non-preferential affirmative action programs are not rendered presumptively unconstitutional.

Part I of this Article defines race-conscious, non-preferential affirmative action as the last wave of affirmative action because it has until now escaped legal challenge, despite vigorous attacks on other forms of affirmative action in recent years. It then analyzes philosophical and legal critiques of affirmative action in general, which vehemently resist the idea that we must look at race if we are ever to get beyond race as a defining feature of American life. These critiques share a fundamental discomfort with the notion that a person's membership in a racial group should have any legal relevance, and exhibit a stubborn unwillingness to endorse any steps that would tend to allow equal access to opportunities between minorities and whites.

Part II contextualizes the Raso v. Lago and Lutheran Church-Missouri Synod v. FCC cases. While these cases arise in very different factual circumstances, both struggle with the definition and implications of "racial classification" and address whether such classifications necessarily burden whites. The two cases exhibit vastly different philosophical underpinnings and interpret
these themes in very different ways, and the legal outcome of each case reflects these stark divergences.

Part III analyzes the maze of Supreme Court jurisprudence that informs the two cases. This discussion includes the Court's treatment of race and racial classifications in both the governmental contracting and redistricting areas. That precedent underscores how Raso and Lutheran Church are "on the edge of developing law," [FN23] and that the ultimate outcome of this issue will have far-reaching ramifications for our legal and practical understandings of the relationship between government and its citizens.

Part IV explains why the Lutheran Church holding fundamentally misunderstands the vital role of non-preferential affirmative action, mistakenly believes that such programs burden whites, and ignores the reality that "to get beyond racism, we must first take account of race." [FN24] Part V shows that the Raso court, by implicitly recognizing the myth of race neutrality, displays a sophisticated understanding of the role race and racism play in limiting access to opportunities. The Raso court crafts a solution that evens the playing field without objectionably burdening non-minorities. The Raso court's approach is correct and should lead developing jurisprudence in this area.

Sharon Foster, [FN1] an African-American civilian employee at the Naval War College in Newport, Rhode Island, wanted to advance her career. [FN2] She "assiduously applied" for more attractive jobs in the Newport Naval Base, [FN3] but her applications were unsuccessful because most of the offices on the base followed a policy of granting preference to internal candidates when promotional opportunities arose. [FN4] Finally, Foster was hired in 1990 as the Professional Affairs Coordinator at the Naval Hospital. [FN5]

 
Shortly after Foster's arrival at the hospital, a new position for a management analyst was created at the hospital. [FN6] Fearful that he could lose the funding for the new position if a protracted applicant search took place, the hospital's Director of Administration, Commander William Travis, decided to undertake a noncompetitive search rather than go through the usual method of recruiting civilian staff. [FN7] Travis directed his staff to cull names from existing files and to make up a list of potential candidates for the new position. [FN8] His staff compiled a list of five candidates, one of whom was Sharon Foster. [FN9] Foster, who was clearly qualified for the position, was the only non-Caucasian on the list and was the only person who was already employed by the hospital. [FN10] Had Travis followed the usual navy yard policy of preferring in-house applicants for promotion--the same policy that had worked against Foster when she had previously submitted applications to other facilities--Foster would have gotten the job. [FN11] But Travis did not follow the hospital's policy.

 
George Warch, the hospital's civilian program specialist, presented Travis with the list of potential candidates. Travis asked why James Berry's name had been omitted. [FN12] James Berry was Warch's "fishing buddy" and an acquaintance of Travis. [FN13] Warch told Travis that Berry was not eligible for the job because of the grade specified for the position. [FN14] Travis directed Warch to rewrite the job description, assigning a lower grade to the job, for which Berry was eligible, and to generate a new list. [FN15] Travis also added computer expertise as a job requirement, a capability that Berry possessed. [FN16] Travis intimated to Warch that, if necessary, he would invoke the Veterans Readjustment Act, which gives veterans preference in certain governmental employment. [FN17] Berry had served in the Navy.

 
With the modified job description, there was only one person on the list-- Berry. Travis named him to the position even though Warch expressed concerns about whether it would look as though they had rigged the result. [FN18]

 
Foster sued the Navy, alleging race discrimination under Title VII of the Civil Rights Act of 1964. [FN19] At trial, the defendants denied that they had discriminated against her; Travis and Warch gave the court "pious assurances that cronyism played no role in Berry's recruitment." [FN20] The lower court rejected these assurances, concluding that the selection of Berry, in the words of the court of appeals, was a "near-classic case of an old boy network in operation, but not a situation in which the employment decision was motivated by racial animus." [FN21]

 
On appeal, the First Circuit panel, which took every opportunity to demonstrate its disapproval of the Navy's actions, affirmed the lower court's judgment. The court of appeals held that it could not overturn the lower court's findings of fact because they were not clearly erroneous. [FN22] Judge Selya concluded that although a fact finder could conclude that Warch and Travis acted out of racially discriminatory animus, the lower court's finding that they were motivated by cronyism rather than discrimination was another permissible inference to be drawn from the evidence. [FN23] Therefore, Sharon Foster, a more qualified black female, lost her employment discrimination suit even though a less qualified white male, who was a friend of the decision maker, got the job.

Compare Sharon Foster's case with that of Sharon Taxman, [FN24] a white teacher in the Piscataway School District in New Jersey. In 1989, the Piscataway School Board accepted the recommendation from the superintendent of schools that it should lay off one of the teachers in the business department at Piscataway High School. [FN25] Normally, the board has no discretion in choosing which employee to lay off because layoff decisions are governed by seniority in accordance with state law. [FN26] In this case, however, there were two employees with equal seniority--they had both started to work on the same day nine years before the layoff. [FN27] One of the teachers, Debra Williams, was black; the other, Sharon Taxman, was white. [FN28]

 
Concluding that Taxman and Williams were equally qualified, the superintendent of schools recommended that the board invoke the district's affirmative action policy in order to make the layoff decision. [FN29] The relevant portion of the affirmative action policy stated: "In all cases, the most qualified candidate will be recommended for appointment. However, when candidates appear to be of equal qualification, candidates meeting the criteria of the affirmative action program [FN30] will be recommended." [FN31]

 
The superintendent of schools made this recommendation because Williams was the only black teacher in the business education department. [FN32] In response, the board independently assessed the classroom performance, evaluations, volunteerism, and certifications of the two teachers and determined that they had equal abilities and qualifications. [FN33] Given this assessment, the board invoked the affirmative action policy to break the tie between the two teachers. Sharon Taxman, the white employee, was laid off. [FN34]

 
Taxman filed a charge with the Equal Employment Opportunity Commission, which in turn brought a Title VII action against the school board. At his deposition, the board president testified that the board had invoked the affirmative action policy in order to provide role models for the students and to promote understanding and tolerance of persons of different backgrounds. [FN35] The federal district court granted Taxman's motion for summary judgment, a judgment that the Third Circuit, sitting en banc, affirmed on appeal. [FN36]

 
The above two cases stand in stark contrast to one another. In Foster v. Dalton, the court approved of the promotion of a less-qualified white male over a better-qualified black female under very suspicious circumstances; in Taxman v. Board of Education, the court invalidated the retention of an equally qualified black female over her white counterpart.

 
Taxman and Foster are troubling cases. Our instincts tell us that the decision made by the Piscataway School Board in Taxman is far less objectionable than that made by Commander Travis in Foster--yet Piscataway's conduct was considered illegal while the Navy's blatant cronyism was immunized. The most obvious reason for the discomfort is the knowledge that in Taxman, the person retained in the position was as qualified as the person who was laid off. The board had to lay off one of the two teachers; educators made the decision that because there were no other blacks in the business department and the students needed role models in all areas, it would be better pedagogically to retain the black teacher. Moreover, the school board did not reach this decision until it had made an exhaustive inquiry, resulting in the board's conclusion that the two teachers were equally qualified. The school board acted cautiously, reasonably, responsibly, and with the educational interests of the students in mind.

 
Our reaction to Commander Travis' decision to promote a "fishing buddy" over Sharon Foster is much more negative. We know that Foster was much better qualified for the position than Berry. In fact, Berry's name did not even appear on the original list; Commander Travis ordered his assistant to rewrite the job description so that Berry's name would appear. Additionally, Foster had applied for other positions at the naval yard but had lost out to internal candidates. This policy of favoring in-house applicants that other departments at the navy yard had invoked against Foster, and to which the hospital usually adhered, was abandoned in order to permit the selection of Berry--a white male. [FN37]

 
The law justifies the disparate results in Foster and Taxman by invoking the principle of race and gender "neutrality" in the decision making process. Under this principle, the law generally prohibits employment determinations based consciously on a person's race or gender. An exception to the "neutrality principle" of Title VII is the doctrine set forth in United Steelworkers of America v. Weber [FN38] and reaffirmed in Johnson v. Transportation Agency of Santa Clara [FN39] permitting race- or gender-based decisions made pursuant to valid voluntary affirmative action plans. [FN40] A valid plan, according to Weber and Johnson, has a purpose that mirrors Title VII's purposes [FN41] and does not unnecessarily trammel the interests of white male employees. Conversely, employment decisions made absent an intent to discriminate because of the employee's protected characteristic are legal.
 

In Taxman, since the board consciously took race into account in making its decision, the court evaluated the Piscataway affirmative action plan under the two prong Weber test and found the plan lacking because it did not have a remedial purpose. [FN42] In Foster, the lower court found that Travis did not consciously consider race or gender; [FN43] for this reason, the necessary element of intent to discriminate was absent. Under the "neutrality principle," as framed by the courts, the Navy did not make an illegal discriminatory decision.
 

But does this form of "neutrality" provide an adequate theoretical basis for the differentiation of these cases? Is there something so inherently wrong with a conscious consideration of race in Taxman that it should determine the outcome of the case? In contrast, why does the absence of a conscious intent to discriminate in Foster automatically relieve the actor of liability? Has discrimination law been misled by a false concept of neutrality and a misshapen notion of preference? Is there an alternative framework that will more appropriately lead to a just resolution of these cases?

 
This Article analyzes these questions and provides a new conceptual framework for approaching the question of race-based and race-neutral decision making. Instead of relying on the simplistic difference between a race-based and race- neutral decision to distinguish between legal and illegal actions, this new framework considers a complex array of issues raised by the decision making process.

 
Part I analyzes the legislative history of Title VII of the Civil Rights Act of 1964 ("Act"), the avowed purposes of the statute, and the tension between concepts of liberty and equality that emerged during the legislative process. This part defines "equality" as the use of merit in making employment decisions; it defines "liberty" as the employer's right to control his workforce by hiring, promoting, and firing whomever he desires. Part I argues that the legislative history, when read in light of the avowed purposes of the Act, is more consistent with a requirement that employers use merit as a basis for making employment decisions, at least when those decisions affect the persons the statute was originally designed to protect: African Americans, other persons of color, and women.

 
Part II demonstrates that recent Supreme Court precedent has unreasonably narrowed the intent requirement in Title VII law, permitting the emergence of the cronyism defense. It shows that the cronyism defense is an exaltation of the employer's liberty interest over an employee's equality right to hiring by merit; it argues that although this narrowing restriction on employers' liberty [FN44] is arguably consistent with the legislative debate, its strict focus and cramped definition of discriminatory intent ignore the extent and type of injury suffered by intended beneficiaries of the Act.

 
This Part shows that recognition of the "merit principle" enunciated in Part I would broaden the definition of intent in discriminatory treatment cases to permit nullification of decisions like those in most of the cronyism cases. It would require an employer who hires on the basis of cronyism to the detriment of one of the intended beneficiaries of the Act to justify the hiring decision on the basis of merit. [FN45]

 
Part III examines the current debate over affirmative action and argues that the perceived dichotomy between affirmative action and merit is false. This Part analyzes the sources of the "antimerit myth" in the case law and popular culture and concludes that the merit principle is consistent with voluntary affirmative action plans as well as with an employer's liberty interests. It links the presence of an invisible white privilege and a precept of black inferiority to an exceedingly narrow definition of merit that judges persons of color and women on the basis of crabbed and rigid criteria defined by the white male norm that fail to consider the overall value of workers. This Part demonstrates merit must be redefined in a broader fashion in order to further the most important goals of the statute. It argues that contrary to popular belief, voluntary affirmative action policies play the vital role of combating current and ongoing discrimination that often results from invisible privilege and unconscious bias. Thus, the antidiscrimination principle is an important justification for voluntary affirmative action and not merely a remedy for egregious wrongs of the past.

 
Part IV demonstrates that the neutrality principle masks the relative harm suffered by victims of discrimination due to "color evasion" [FN46] and "power evasion." [FN47] Using a set of five hypothetical situations, it illustrates the types and qualities of injuries suffered by unsuccessful job applicants in different historical and social contexts. These hypotheticals demonstrate that although the existence of a race-based or neutral decision making process may affect the victim's injury, the extent and type of injury is much more inextricably related to the historical, social, and economic context in which the employment decision is made. Part IV concludes that the Act's focus on discriminatory intent as a measure of liability is simultaneously underinclusive and overbroad because it ignores the very real differences between the types of injuries suffered by employees in different historical, social, and economic contexts.

 
Part V proposes a conceptual framework to decide Title VII cases that takes into account a complex array of issues. These issues include the historical legacy of slavery and discrimination against African Americans; the presence of invisible white privilege; the precept of black inferiority; the existence of stereotypes; conscious and unconscious biases; the tension between the law's approach to the values of equality (a right to merit-based job decisions) and liberty; and the types of injuries suffered by different "victims" of discrimination.
 

To its critics, one of the flaws of race-based affirmative action is that its main beneficiaries are economically privileged members of the eligible minority groups. [FN1] Supporters of race-based affirmative action, particularly in the sphere of education, have responded by claiming (implicitly or explicitly) that economic inequality is not, in fact, the reason for race-based affirmative action at all. Instead, they embrace diversity as affirmative action's central goal--and in so doing, they find a justification for continuing to include the middle class in minority affirmative action programs. [FN2] The decision to concede away the economic inequality rationale for including the minority middle class in affirmative action programs has the appearance of being relatively cost-free, given that the Supreme Court has rejected "societal discrimination" as a ground for affirmative action. If the Court will not buy it, it seems to make perfect sense to give it away.

 
But appearances are deceiving. This paper shows, in Part I, that the cost of conceding away the economic disadvantage of the minority middle class is in fact unacceptably high. [FN3] Although the Supreme Court's affirmative action jurisprudence prefers diversity to societal discrimination as a rationale for affirmative action, [FN4] the diversity rationale is highly problematic and ought not to be made to stand alone. Instead, the diversity rationale is most persuasive when it is augmented by the view that past and present race-based economic inequality is the reason we cannot achieve meaningful levels of integration without using affirmative action. Second, this paper shows, in Part II, that the minority middle class--not merely the minority poor and working class--suffers race-based economic inequality. Using the black middle class as my example, [FN5] I demonstrate that the black middle class is systematically worse off than the white middle class on numerous economic dimensions. I argue in Part III that so long as race-based affirmative action remains an available strategy, there is good reason to use it to redress the economic gap between the white and black middle classes. I conclude in Part IV by admitting and examining some of the political costs of relying on the relative economic disadvantage of the minority middle class as a rationale for affirmative action.

 
I shall contend, in Part I.B., that one of the key reasons contemporary supporters of affirmative action heavily rely upon the diversity rationale for affirmative action is the awkwardness of defending affirmative action for the black middle class on any other ground. But there is, of course, a far simpler explanation for the popularity of the diversity rationale: the economic-inequality rationale for affirmative action rests upon the existence of present and past "societal discrimination." Anyone who has read the Supreme Court cases knows that the Court does not accept the remedying of past or present societal discrimination as an acceptable justification for affirmative action. [FN6] Why, then, look any further for an explanation of the shift to diversity?

 
To answer this question, I must begin by explaining why the Supreme Court's decisions are not a complete explanation for the current emphasis on diversity as the justification for affirmative action. And, since I do not support dismantling affirmative action, [FN7] I must also explain why, in light of the Court's rejection of societal discrimination as a rationale for affirmative action, it is neither legally irresponsible to continue to stress the existence of societal discrimination nor politically irresponsible to claim, as I do here, that the diversity rationale is unconvincing unless it is coupled with an understanding that race-based economic inequality stands in the way of achieving diversity without affirmative action. Part I.A. addresses both issues.

 
A. Diversity Versus Inequality: Why the Court Is Not the Sole Explanation or Constraint

 
For decades, the Supreme Court was the major institutional locus for the debate on affirmative action. This does not mean other institutions, private and public, were not and are not involved in affirmative action decisionmaking. To the contrary, the Supreme Court has never adopted the view that affirmative action is constitutionally required in those situations in which it is constitutionally or, under Title VII, statutorily permitted. [FN8] Thus, public and private institutions have for years acted under the assumption that they have complete freedom to refrain from using affirmative action, and their decisionmaking has thus always been a part of the affirmative action debate. [FN9] State polities have begun to use the initiative process to amend state constitutions to bar affirmative action and have met with political and legal success. [FN10] Congress remains an important actor, not only because it controls the use of affirmative action in the federal sector, but also because Congress has long assumed itself free to order private employers to cease practicing affirmative action by amending Title VII and similar statutes to provide that affirmative action is actionable discrimination. [FN11] Affirmative action thus has been a project of the American polity at large, not of the Supreme Court alone.

 
Why, then, was the Supreme Court so central to the affirmative action debate? Power is, of course, the most important reason. Congress and the executive can and should consider the Constitution in their decisionmaking, but a ruling by the Court that an action of the political branches is unconstitutional ends the debate. [FN12] Because the constitutionality of an action may depend upon its purposes, it is important that those who create and implement affirmative action programs correctly anticipate which purposes the Court would uphold and which it would strike down. Proponents of affirmative action have every incentive to rely solely upon justifications that they know will pass judicial muster. Thus, the Court has the unique power to control not only what institutions do on the affirmative action front, but also what they say--and thus the Court directly affects the arguments the public will hear from government officials.

 
In statutory interpretation, the formal powers of Congress and the Court are reversed: the Court can interpret an ambiguous statute to prohibit affirmative action, but Congress is always free to restore the status quo by adopting new legislation. [FN13] But legislative overrides are easier to threaten than to deliver, given the difficulty of turning the congressional agenda to any particular issue. [FN14] Furthermore, the Supreme Court itself changes the playing field when it steps up to bat. For example, in United Steelworkers of America v. Weber, [FN15] the Supreme Court interpreted Title VII to permit voluntary affirmative action because it found that Congress in 1964 could not possibly have intended anything so retrograde as the principle of complete color blindness. This decision bolstered the argument that antidiscrimination without affirmative action was unthinkable. [FN16] And, most important, the Court's approval of voluntary affirmative action made affirmative action the status quo, placing the burden on opponents of affirmative action to defend the need for change. [FN17]
 

But sheer power is not the only reason that the Court has dominated the affirmative action debate for so long. Another reason is publicity. The decisions of Congress to mandate affirmative action (such as minority set- asides in government contracting), of administrative agencies to use affirmative action as their antidiscrimination mechanism of choice, or of private businesses to adopt affirmative action plans took place in relative secrecy. Affirmative action provisions spread from statute to statute with little public attention. [FN18] The Equal Employment Opportunity Commission's reasons for favoring affirmative action are now the subject of historical discussion, [FN19] but they were not laid out before the public in clear and resonant language. Private businesses are private, and so long as the EEOC was on their side, they could work out their programs in the seclusion of their own headquarters or at the negotiating table with their own unions. None of these chapters of the affirmative action story were subject to the same scrutiny that the Court's divided pronouncements were; only in the case of the courts (most prominently the Supreme Court) do individual governmental actors clearly articulate the reasons for their decisions in a form that is made accessible to the lay public by the news media. [FN20]

 
The Court's constitutional affirmative action decisions suggest a continuum of the relative safety of potential justifications for affirmative action programs. [FN21] At the safest end of the continuum is the use of affirmative action to remedy the institution's own past discrimination. [FN22] But past discrimination is notoriously difficult to prove, and may in fact be absent. Under Regents of the University of California v. Bakke, [FN23] colleges and universities may use affirmative action to achieve diversity in admissions: the Fifth Circuit in Hopwood v. Texas may have declared Bakke dead, [FN24] but it lacked the power to make it so. Whether diversity is a proper rationale for affirmative action beyond the specific context of Bakke is an open question. [FN25] "Diversity" as a broad-based rationale thus occupies an intermediate position on the affirmative action safety continuum. The most dangerous end of the continuum is occupied by "role model" theories [FN26] and by "societal discrimination." [FN27] The cautious position, then, is to abandon societal discrimination as a rationale for affirmative action and to justify affirmative action entirely on diversity. Diversity is not a sure thing, but it seems to beat the practical alternatives.
 

But is it true "caution" to refrain from arguing that societal discrimination exists, even when arguing to the Court? To restrict oneself to arguments the Court now accepts is to forget the basic legal realist tenet that the reason a judge votes the way he votes is not necessarily the reason he is prepared to set forth as his official rationale. [FN28] A judge will be more likely to read precedent as permitting a broader range of action if the judge is personally convinced there are good reasons to do so, even if those good reasons are reasons (like societal discrimination) that must go unstated. Thus, a justice faced with the question whether diversity as a justification for affirmative action survives strict scrutiny might well be influenced by her (unstated) views about why diversity cannot be achieved without affirmative action--which might well turn on the effects of societal discrimination. [FN29] Ultimately the ideal is to convince the Court to bring its stated rationale into line with its felt commitments, and to do so where necessary by reconsidering its prior opinions and decisions. Until that happens, the advocate has two jobs: to convince the Court it wants to go her way and that there is an intellectually honest way to do so. To the extent that advocates limit their arguments to ones that are currently acceptable to the Court as permissible rationales, they do only part of their job.

 
Suppose, however, that we stop being judicial realists for the moment and take the view that, in advocacy to the courts, one cannot allude to arguments the courts have rejected in the past. Does that mean that public discourse on affirmative action should restrict itself to rationales upon which the courts are presently willing to rely? Absolutely not.

 
It is more true now than ever that the judiciary is just one of many powerful loci of the affirmative action debate. Public institutions are taking advantage of their power to refrain from using affirmative action (as the Regents of the University of California did in ordering its constituent institutions to refrain from affirmative action), and state and local governments are asserting their power to restrict the use of affirmative action by public institutions within their own jurisdictions (as the citizens of California did in Proposition 209, the California Civil Rights Initiative). Affirmative action is thus in danger not just as constitutional doctrine, but as public and industrial policy. As a consequence, affirmative action advocates must convince a wide range of public and private decisionmakers that affirmative action is a good idea. What the Court finds convincing may not convince the broader polity, and vice versa. It is thus a good thing that academics and publicists are not barred from arguing that societal discrimination is also a reason to retain affirmative action, even if public bodies could not satisfy the Court by claiming that the government's "purpose" in pursuing affirmative action is the eradication of societal discrimination. [FN30]
 

B. Diversity and the Black Middle Class

 
We have seen that the Supreme Court shapes the public debate on affirmative action but by no means completely constrains it. If the defense of affirmative action has come to center on the value of diversity, the Supreme Court's jurisprudence is not the only reason. The other major reason is that the diversity rationale seems, at first glance, the only way to justify affirmative action once it is acknowledged that its beneficiaries are generally the most economically privileged members of the eligible minority groups.

 
When William Julius Wilson pointed out that race-based affirmative action benefits the black middle class and not the "truly disadvantaged" members of the black community, [FN31] he may not have realized the political implications of his work. He probably did not intend that his scholarship would become central to those arguing for the abandonment of existing affirmative action programs. [FN32] It was instead his hope that his work would support political action to provide a different kind of intervention for the poorest blacks-- namely, the kind of massively expensive anti-poverty programs that would be politically palatable only if they reached out to all poor people without regard to race. As of yet, Wilson's work has not borne the fruit he hoped for; we have not achieved massive intervention on behalf of the poor. [FN33] Wilson's work has, however, borne unexpected fruit for the anti- affirmative action camp, which uses it to argue that race-based affirmative action ought to be dismantled because it helps privileged blacks, not disadvantaged blacks.

 
Wilson's work calls for an answer from advocates of race-based affirmative action. No one, to my knowledge, has successfully refuted Wilson's claim that the African Americans who succeed in taking advantage of affirmative action are economically advantaged compared to the African Americans who do not. Nor has anyone succeeded in refuting the claim that there are whites who are far poorer than the average black affirmative action beneficiary. [FN34] The best one can do in refuting Wilson is to argue that affirmative action was never meant to be a program for the poor, and need not be defended as one. But that only raises more questions: What is the reason for affirmative action? Why should so much energy be spent protecting a program that mostly helps relatively privileged people?

 
From that point, the argument can go in two directions. One is to justify affirmative action on grounds other than economic disadvantage. The other is to admit that affirmative action is about economic disadvantage but to claim that the relevant economic comparison is between the black middle class and the white middle class, rather than between the black middle class and either the white or the black poor--and, further, that the relative disadvantage of the black middle class is a serious enough problem to justify the use of affirmative action. I advocate the latter strategy, but it is the former that is now in the ascendancy.
 

Affirmative action in law faculty hiring continues to provoke controversy. In this Article, Professors Merritt and Reskin seek to inform the ongoing debate by reporting results from the first comprehensive empirical study of the effects of sex and race on tenure-track hiring at accredited law schools. Their analysis controls for educational credentials, work experience, job search strategies, and a number of personal characteristics. By using all of these factors, Professors Merritt and Reskin evaluate the independent effect of each factor on the prestige of the school where a tenure-track professor was hired, on the initial academic rank of the new professor, and on the courses the new professor taught.

 
Their analysis indicates that white women and men of color, but surprisingly not women of color, began teaching at somewhat more prestigious schools than white men with comparable credentials. The advantage, however, was relatively modest. Educational credentials and work experience accounted for the bulk of explained variance in the prestige of schools where tenure-track professors were hired, and even job search strategies and personal characteristics such as age or marital status accounted for three times more of the explained variance than did sex and race. Furthermore, the results suggest that sex bias affected other measures of job status. Men were more likely than women to begin teaching at a higher professorial rank. In addition, men were more likely than women to teach constitutional law, a high-status course that can enhance a professor's career prospects. On the other hand, women were more likely than men to teach trusts and estates or skills courses, both lower status subjects that may diminish a professor's career prospects.
 
 

Introduction ............................................. 200
  I.  Methodology .............................................. 206
        A. Population .......................................... 207
        B. Data Sources ........................................ 210
        C. Dependent Variables ................................. 211
          1. Institutional Prestige ............................ 211
          2. Initial Tenure-Track Rank ......................... 215
          3. Courses Taught .................................... 216
        D. Independent and Control Variables ................... 220
        E. Omitted Variables ................................... 225
 
 II.  Results .................................................. 230
        A. Institutional Prestige .............................. 236
          1. The Top Sixteen Law Schools ....................... 236
          2. Institutional Prestige as an Ordinal Variable ..... 246
        B. Rank of Initial Tenure-Track Position ............... 252
        C. Course Assignments .................................. 258
          1. Constitutional Law ................................ 259
          2. Skills Courses .................................... 261
          3. Trusts and Estates ................................ 263
          4. Family Law......................................... 264
          5. Corporations ...................................... 266
        6. Taxation .......................................... 266
          7. Institutional Decision or Personal Choice? ........ 267

  III.  Discussion ............................................... 274
        A. Overall Effects of Sex, Race, and Credentials ....... 274
          1. Sex and Race ...................................... 274
          2. Credentials ....................................... 275

        B. Lessons for the Affirmative Action Debate ........... 280
          1. The Modest Size of Race and Sex Preferences ....... 280
          2. Fit Between Means and Ends ........................ 282
          3. Ongoing Sex Discrimination ........................ 289
          4. Intersection of Sex and Race ...................... 290
          5. Affirmative Action Myths and Truths ............... 291
        C. Beyond Credentials: Directions for Future Research .. 294

      Conclusion ............................................... 299

      Appendix 1 ............................................... 301

      Appendix 2 ............................................... 303

 
Sex and race are hot topics in law faculty hiring. Some professors charge that discrimination continues to shackle women of color, white women, and men of color, preventing them from securing tenure-track positions at top law schools. Derrick Bell provoked national headlines when he accused Harvard Law School of discriminating against women of color in hiring and took an unpaid leave of absence to register his discontent. [FN1] Judge A. Leon Higginbotham similarly protested the paucity of women and minorities at the University of Chicago School of Law by declining to judge a moot court competition at that institution. [FN2] Numerous legal scholars have echoed these challenges, claiming that sex and race biases still distort the hiring process in American law schools. [FN3]

 
At the same time, an equally vociferous group of professors complains that law faculty hiring is biased unfairly in favor of women and minorities. Several white men recently voiced this perspective by publishing narratives intimating that law schools refused to consider them for faculty appointments--or substantially discounted their credentials--because of their race and sex. [FN4] The Dean of Stanford's Law School jokingly evoked the same sentiment by declaring: 'Copr.[I]t's not a good day for white males. I'm glad I have my job here." [FN5] And a respondent to one of our own surveys more caustically registered his concern by identifying his 'race' as 'white male' and adding: 'That's why I don't have a better job.' [FN6]

 
This tension in faculty lounges mirrors a broader social debate over affirmative action, discrimination, and the current role of sex and race in the job market. [FN7] A rising chorus of scholars, politicians, and members of the public attacks affirmative action as a spoils system for white women, men of color, and women of color. [FN8] These critics complain that affirmative action programs allow women and minorities to steal jobs from better qualified white men. The results, they claim, are unfair, undemocratic, socially divisive, and unresponsive to the most harmful effects of racism. [FN9]

 
Other social commentators point to evidence of ongoing discrimination against women and minorities in education, housing, and the job market, contending that affirmative action is necessary to remedy those evils. [FN10] Without explicit attention to sex and race in hiring, these advocates maintain, unconscious biases and 'old boy' networks will continue to exclude women of color, white women, and men of color from high-paying jobs--even when these nontraditional applicants are as qualified as the white men who obtain the positions. [FN11]

 
Social scientists have begun to inform this tempestuous debate by tracing the impact of affirmative action on both targeted groups (women and minorities) and other workers. [FN12] We build upon that literature by offering the first full-scale, multivariate analysis of the effects of sex and race on tenure- track hiring at accredited law schools. [FN13] Our study focuses on professors who began their first tenure-track position at an accredited law school between the fall of 1986 and the spring of 1991. We found little statistically significant evidence that law schools preferred white women, women of color, or men of color over white men in this population. We identified a modest preference for white women and minority men on just one of nine job characteristics we explored: after controlling for academic credentials, work experience, and personal characteristics, white women and men of color obtained tenure-track jobs at slightly more prestigious institutions than did white men with equivalent credentials. The preferences, however, were small and did not play any special role at the top sixteen law schools. Most surprising, these modest preferences conferred no advantage on women of color; those women fared no better than comparably qualified white men in obtaining jobs at prestigious institutions.

 
Equally important, we identified evidence of persistent sex bias in law faculty hiring. Even after controlling for credentials and other relevant characteristics, law schools hired men at significantly higher professorial ranks than they hired women. Likewise, men were significantly more likely than women with equivalent backgrounds to teach constitutional law, a high-status course preferred by many new professors. Women, on the other hand, were significantly more likely than men with identical credentials to teach both skills courses and trusts and estates--lower status offerings that provide fewer opportunities for career advancement. Both white women and women of color suffered from these starting-rank and teaching-assignment disadvantages. These findings suggest that, even with the presence of affirmative action programs, law schools marginalize women on their faculties.

 
In Part I of this Article, we explain both our methods and why we focused our study on tenure-track law professors. We then outline our empirical findings in Part II, identifying sex and race differences in the ability of professors to win tenure-track appointments at the most prestigious institutions, secure jobs at the highest ranks, and obtain assignments teaching particular courses. Part III then discusses the import of these findings for the ongoing debate over affirmative action in both law schools and other workplaces. [FN14]
 

 
A precondition for any discussion of affirmative action is defining the meaning of the term. The concept of affirmative action has been bandied about in such an elastic way that many people view it as a code word for reverse discrimination, lower standards or rigid quotas. As used in this paper, affirmative action is a flexible tool to promote equality of opportunity in the employment context. The purpose of affirmative action is to remedy past and present discrimination, as well as to prevent future discrimination. Consistent with the Affirmative Action Guidelines promulgated by the Equal Employment Opportunity Commission, [FN1] an affirmative action plan identifies employment policies and practices which present barriers to the hiring, advancement and retention of women and minorities, and establishes goals and timetables as a device for measuring progress in overcoming racial and gender discriminatory practices.
 

An affirmative action plan is based on merit, not the lowering of standards. An affirmative action plan analyzes the employer's current workforce to determine whether there is a manifest imbalance between the racial and gender composition of its workforce and the composition of the "qualified" labor pool from which the employer draws its employees. Where such an imbalance occurs, the employer analyzes its job policies and procedures in an attempt to identify, and discontinue, those practices which may be causing the imbalance. The plan establishes a timetable, taking into account employee turnover and the legitimate interests of non-minority employees, during which time the employer can expect to achieve a racial- and gender-balanced workforce. The plan does not result in hiring unqualified applicants, nor does it completely bar non-minority candidates from consideration.

 
This paper will discuss affirmative action only as it is practiced in the private sector employment context. Affirmative action undertaken by federal, state or local government entities is subject to a different set of legal restrictions, and may be justified by different considerations, which will not be discussed herein.

 
The legality of the implementation and maintenance of voluntary private sector affirmative action plans under Title VII of the Civil Rights Act has been affirmed by the Supreme Court in both United Steelworkers of America v. Weber, 443 U.S. 193 (1979) and Johnson v. Transportation Agency, 480 U.S. 616 (1987).

 
In Weber, the employer's affirmative action plan set aside 50% of the openings in a craft-training program for black employees, to be effective until the number of black craft workers at the plant approximated the percentage of blacks in the local labor force. Less than 2% of the employer's craft workers were black, even though 39% of the labor force was black. The employer established the craft-training program in order to train its own production workers to fill craft openings. No special qualifications were needed to enter the program; the employer used seniority and the affirmative action plan to designate the entrants. A white production employee who was denied entry to the program challenged the affirmative action plan in court, alleging that it discriminated on the basis of race in violation of Title VII. The lower courts agreed, holding that all employment preferences based on race, even those pursuant to an affirmative action program, violated Title VII. The Supreme Court disagreed and reversed the holdings of the lower courts.

Initially, the Court noted that an interpretation of Title VII to forbid all race-conscious affirmative action would be contrary to the purpose sought to be achieved by the law. The statute was intended to cause employers to evaluate their employment practices and attempt to eliminate "the last vestiges of an unfortunate and ignominious page in this country's history." [FN2] Title VII was not meant as a purely reactionary statute for prosecuting offenders, but was also intended to spur proactive conduct by employers aimed at preventing discrimination.

 
In discussing the Congressional purpose behind Title VII, the Court read the language of § 703(j) as an indication that Congress chose not to prohibit all race-conscious affirmative action. Section 703(j) states that nothing contained in Title VII "shall be interpreted to require any employer . . . to grant preferential treatment . . . to any group because of . . . race . . . ." This language does not, however, forbid voluntary action by an employer.

 
Additionally, the Court noted that Congress wanted to prevent undue federal regulation of business which would interfere with or limit "traditional business freedom" [FN3] or "management prerogatives" [FN4] in running the business, including the freedom to establish voluntary affirmative action programs.
 

Finally, the Court suggested several criteria to consider in determining whether an affirmative action plan is "bona fide" in the sense that it is consistent with the policy and purpose of Title VII. [FN5] First, the plan must be designed to break historic patterns of racial segregation in employment opportunities and jobs. In Weber, the clear imbalance between the racial composition of the employer's craft force and the local labor force suggested such historic patterns. Second, the plan must not unnecessarily trammel the interests of white employees. In Weber, no white employees lost their jobs because of the plan, and operation of the plan did not absolutely bar white employees from entry into the training program (indeed, 50% of the slots were filled by whites). Last, the plan must be a temporary measure designed to eliminate racial barriers to employment, not to maintain an already achieved racial balance. In Weber, the plan itself stated that it would end when the percentage of black craft workers approximated the percentage of blacks in the labor force.
 

In Johnson, the employer used an affirmative action plan in making promotions to jobs in which women were significantly under-represented. [FN6] The plan required the employer to consider gender as a plus factor when making a decision about whom to promote from among a pool of qualified applicants. When a job vacancy arose for a position of dispatcher, twelve employees applied for the promotion, and seven were found to meet the qualifications for the job. In making its choice from among the seven qualified candidates, the employer considered the fact that there were no women currently employed in the skilled craft category and promoted a woman. A man who was rejected for the job alleged that the employer's decision was sex discrimination in violation of Title VII. The Court of Appeals held that the employer had acted pursuant to a bona fide affirmative action plan; therefore, taking gender into account was lawful under Title VII.

 
The Supreme Court agreed, and in its holding reaffirmed the crucial role that voluntary employer action plays in eliminating the effects of discrimination in the workplace. The Court found that the employer's affirmative action plan met the criteria established in Weber and was therefore valid under Title VII. There was a manifest imbalance between the percentage of qualified women in the labor force and the percentage of women actually employed by the employer. The plan did not authorize the absolute promotion of women; rather, gender constituted one factor among others, including qualifications, to be taken into account in making the decision. No male employee lost his job due to the plan, nor was any male employee barred from consideration for a promotion. Lastly, the plan was temporary in nature, designed to attain a balanced workforce.
 

The force and continuing applicability of the Weber and Johnson decisions to private sector affirmative action plans has not been undermined by the Supreme Court's recent rulings on the use of set-asides and minority preferences in the public sector. The latter cases, including the Court's most recent decision in Adarand Constructors, Inc. v. PeDna, 115 S. Ct. 2097 (1995), are firmly embedded in constitutional jurisprudence and equal protection analysis and apply the strict scrutiny test to governmental conduct based on race. Such constitutional analysis does not, however, apply to non- governmental, private sector employers. Thus, the decisions in Weber and Johnson, which not only uphold the legality of voluntary private sector programs, but also emphasize their consistency with the underlying goals of Title VII, are as persuasive and binding today as when they were issued.
 

Introduction
 

I. Defining the Problem
A. Are Blacks Underrepresented in Corporate Firms?
B. Why Study Institutions and Incentives?
1. Differential Abilities
2. Lack of Interest
3. Racism
4. Preaching to the Unconverted
C. Affirmative Action
 
II. The Model: Monitoring Among the Human Capitalists
A. Discrimination in High-Level Jobs
1. Subjectivity and Monitoring
2. When It's Cheaper to Overpay
3. Tournament Theory
4. "Efficient" Discrimination
B. Making Elite Law Firms
1. Good Lawyers are Made, Not Born
2. An Efficient Model in an Age of Opportunism
a. High Salaries
b. Pyramiding
c. Tracking and Training
 

III. The Application: Mapping The Racial Law of Averages
A. Recruitment
1. The Process
2. Signals and Signaling: Stocking the Pipeline and Protecting the Franchise
3. Race and Recruiting
B. Retention, Promotion, and Survival
1. Monitoring, Mentoring, and Marketing: Getting on the Training Track
2. Concrete Ceilings and Slippery Floors: The Black Experience in Corporate Firms
a. Mentoring and Irrationality
b. Visibility and Tokenism
c. Bringing the Outside In
d. Rational Strategies in the Face of Reasonable Fear
i. Low-Risk Strategies
ii. High-Risk Strategies: The Litigation Trap
e. The Revolving Door
f. Unequal Partners
i. The External Market
ii. The Internal Market: Getting the Franchise

 
IV. The Solution(s)?: Finding Efficient Responses to Efficient Discrimination
A. Anti-discrimination Law
B. Institutional Reform
C. Diversity Training
D. Stimulating Demand
E. Affirmative Action
1. Standards
2. Effort: Lowering the Price of the Ticket to the Tournament
3 Stigma

Conclusion: Choosing a New Path
 
Appendix

 
Although the number of black students graduating from law schools has increased significantly in recent decades, blacks still make up a very small minority of the lawyers working in large corporate law firms. Available data indicate that these firms hire few blacks, and that those they do hire are more likely than their white peers to leave the firms before becoming partners. Conventional explanations blame the underrepresentation of blacks in corporate firms on either the racism of firms and their clients, or a shortage of qualified, interested black candidates. While acknowledging that in some instances these factors may help to explain the problem, this Article looks behind them to examine institutional factors that tend to perpetuate the existing underrepresentation. Specifically, the Article shows how the ways in which large corporate firms recruit and train lawyers tend both to shield discriminatory choices between black and white candidates from any competitive disadvantage, and to discourage black law students and lawyers from investing in skills that will enable them to succeed within corporate firms. Thus, the Article argues, firms' hiring and training decisions both shape and are shaped by the strategic choices of black candidates, with the net effect of keeping all but a few blacks from being hired and succeeding in the firm setting. Finally, this Article explores the implications of these incentives for five commonly proposed tools for diversifying corporate law firms: anti- discrimination laws, race-neutral institutional reforms, diversity education within firms, demand-creation initiatives, and supply-side initiatives to encourage hiring and promotion of black lawyers.

This Article addresses what for many is an uncomfortable reality: Despite a substantial increase in the number of black students attending law school over the last forty years, African Americans still constitute only a tiny percentage of the associates and partners working in the nation's largest corporate law firms. [FN1] Given the legal profession's role in championing the principles of non-discrimination and equality of opportunity, this reality is particularly troubling. More generally, however, the fact that blacks have had little success breaking into the upper echelons of the elite bar is emblematic of a deeper and more intractable set of problems facing those interested in workplace integration.

 
Forty years after the Supreme Court's landmark decision in Brown v. Board of Education, [FN2] society has made substantial progress toward eradicating the kind of overtly racist policies that excluded blacks from virtually every desirable sector of the economy. For many blacks, these changes have produced a dramatic growth in income and opportunity. In recent years, however, it has become painfully clear that simply dismantling America's version of apartheid has not produced economic parity between blacks and whites. [FN3] Although poor blacks have benefited the least from the civil rights revolution, "high level" jobs in business and the professions have also proved surprisingly resistant to change. [FN4] The fact that blacks have made so little progress in breaking into the corporate law firm elite--particularly at the partnership level--fits this larger pattern.

 
Commentators generally offer one of two explanations for this "glass ceiling" effect. The first, generally proffered by firms, posits a shortage of black applicants with both the qualifications and the interest necessary to succeed in the demanding world of elite corporate practice. [FN5] The second, most often articulated by blacks, blames the slow progress on continued racism both inside corporate firms and among the clients upon whom these entities depend for their livelihood. [FN6]
 

As we argue below, both the "pool problem" and continuing racism against blacks play important roles in determining the employment opportunities available to African American lawyers. Standing alone, however, each explanation begs important questions. The "pool problem" explanation begs the question of whether the existing hiring and promotion criteria utilized by elite law firms to determine who is in the pool fairly and accurately predict future productivity. The racism story, on the other hand, fails to explain why firms that discriminate by refusing to hire or promote qualified black lawyers do not suffer a competitive disadvantage when those workers are employed by their competitors.

 
In order to arrive at a more thorough understanding, we must move beyond this familiar dichotomy. We do so by taking a closer look than either of the standard explanations at how corporate firms structure, and are structured by, the relevant markets for labor and clients. [FN7] Our interest in this neglected institutional dimension is the product of our prior work on race, professionalism, and markets. One of us is engaged in the ongoing study of the legal profession with particular attention to the experiences of black lawyers. [FN8] The other is studying how particular market conditions allow firms to insulate some kinds of discriminatory decisionmaking from the disciplining effects of competition. [FN9] In this Article, we seek to combine these two perspectives by offering a preliminary account of how corporate firms recruit and retain lawyers and why these practices may adversely affect the employment prospects of black lawyers. [FN10]

 
Our argument proceeds in five parts. Part I briefly summarizes the available data on blacks in corporate law practice and defends our claim that an institutional analysis is a necessary component of any plausible explanation for these numbers. We also set out our reasons for initially bracketing (to the extent possible) the impact of affirmative action on law firm hiring and promotion practices. We return to this issue in Part IV.
 

In Part II, we present a stylized model of the contemporary elite corporate law firm. The model is premised on two related features of professional work: the inherent subjectivity of quality assessments and the difficulty and expense of monitoring. In response to these realities, we posit that it is efficient for firms to adopt the following tripartite strategy: high wages to create a large pool of available workers and to motivate those lawyers who are hired to work with relatively little supervision; a high associate-to-partner ratio, thus further encouraging associates to work hard in the hopes of becoming partners while at the same time allowing the firm to spread legal work among many lawyers with varying levels of knowledge and skill at the lowest possible cost; and a tracking system whereby the pool of associates is divided into those who will receive scarce training resources and those who will work on relatively undemanding assignments.

 
These three institutional features of contemporary elite firms, we assert, disproportionately disadvantage black lawyers. [FN11] Two tendencies contribute to this result. First, because firms hire a large number of associates from a pool that has been artificially inflated by high salaries and ask many of them to do relatively undemanding work, these institutions have little incentive to invest in obtaining detailed information about the quality of potential employees. Hence, individuals within the firm can use race as a factor in their decisionmaking without hurting the firm's bottom line. The same goes for retention and promotion. Decisions to invest scarce training resources in average whites as opposed to average blacks will not hurt the firm's chances of producing the small number of high quality partners that it needs to guarantee its productivity in future years. As a result, firms have little incentive to root out employment decisions that, either consciously or unconsciously, prejudice blacks or favor whites.

 
Second, because firms have no incentive to stop these practices, black lawyers in firms (as well as those contemplating joining firms) are more likely to choose human capital strategies that, paradoxically, decrease their overall chances of success in these environments. Since blacks reasonably believe that they face an increased risk that their abilities will be unfairly devalued or overlooked, they have an incentive to overinvest either in avoiding visible negative signals or in obtaining easily observable positive signals that clearly identify them as superstars. Both of these strategies, however, are potentially counterproductive to the extent that they diminish a black lawyer's opportunity or incentive to obtain the skills upon which success at the corporate law firm ultimately depends.

 
In Part III, we test our model against the limited amount of currently available empirical and anecdotal evidence on black corporate lawyers and our own preliminary research on black Harvard Law School graduates. This data is far too tenuous and incomplete to prove definitively the institutional dynamic we posit. Nevertheless, what information there is points in the direction predicted by our model. Although these institutions spend a considerable amount of time and money on recruiting, firms rely on a small number of easily visible and rankable criteria at the initial screening phase, while at the same time relying on subjective judgments about "personality" and "fit" (as opposed to other easily obtainable indicia of quality) during the less visible parts of the process. This combination of objective and subjective decision-making disadvantages black applicants by falsely conveying the impression that the visible and rankable criteria that firms rely on are tightly correlated with quality, while simultaneously allowing individuals within firms to discount a black candidate's signals based on subjective judgments about personality and fit. Similarly, when selecting associates to mentor and train, partners rely on a combination of a few objective signals and their own subjective judgments in order to determine which associates are likely to have the kind of viable long- term careers with the firm that make them good training prospects. Once again, the data suggests that blacks are less likely to be selected to receive this essential good. Finally, as our model predicts, black students and associates often react to these heightened barriers by choosing human capital strategies that further diminish their chances for success in this environment.
 

Part IV examines the implications of our analysis for five of the most commonly proposed solutions for diversifying corporate law firms: anti- discrimination laws such as Title VII, race-neutral institutional reforms such as formal training programs and associate review, educational initiatives such as diversity training, demand-creation initiatives designed to encourage corporate clients to hire black lawyers, and supply-side initiatives that encourage firms to make affirmative efforts to increase the number of blacks they hire and promote. Each of these mechanisms has the potential to improve oppotunities for black lawyers. The structural characteristics of the elite firms described in Part II, however, make it difficult for this potential to be realized.
 

Part V therefore concludes by briefly examing why elite firms might change these strucural impediements. Although we claim the existing structure is efficient, the argument that it is the only efficient response to the problems of subjectivity and monitoring ignores both historical contingencies that produced these practices and the unprecidented volitility of the market for elite firm services. Ironically, we believe that the sense of crisis that currently pervades not only elite firms, but the enitre legal profession, offers a unique opportunity to chart a new path that will enable there institutions to move closer to the ideals of fairness and equality that constitute the best part of this country's legal tradition.
 

The employment patterns of "nontraditional" [FN1] workers in the United States show two conflicting characteristics. On the one hand, researchers have observed a continuing increase in the rate of participation of nontraditional workers at multiple levels in the work force. For example, the proportion of women white collar workers increased from twenty-two percent in the late 1960s to forty-six percent in 1992. [FN2] Similarly, the average job tenure for nontraditional workers has also increased. [FN3] For example, although males in the thirty-five to forty-four year old age group have experienced a small decline in job tenure, women in the same group have seen increasing participation in the high tenure categories. [FN4]

 
On the other hand, nontraditional workers have also experienced the phenomenon of the "glass ceiling." [FN5] This glass ceiling is a barrier that prevents women and minorities from moving to high level managerial positions of power and authority within organizations. [FN6] The glass ceiling has recently been documented as existing at a lower level in organizations than originally thought. [FN7] One recent study of Fortune 1000 companies indicated that although about thirty-seven percent of employees are women and about fifteen percent are minorities, only seventeen percent of all managers are women, and only six percent of all managers are minorities. [FN8] At executive level managerial positions the discrepancies are even larger: only about six percent are women, and only three percent are minorities. [FN9]
 

The advancement of women and minorities in the work force, as illustrated by their increased labor force participation numbers, has in part been attributed to the protection provided under Title VII. [FN10] Title VII has been fairly successful in eliminating barriers to entry into some positions previously closed to women and minorities. [FN11] Early efforts at combatting discrimination were aimed at lower level, often entry level, jobs, because they represented the greatest number of employment opportunities and had the smallest number of requirements for qualification, including education and experience. [FN12] In particular, Title VII has helped reduce barriers that prevent women and minorities from entering nontraditional areas of employment.

 
If Title VII has been successful at the entry level, we might expect it to be similarly successful in protecting women's and minorities' interest in advancement to higher levels of employment. Instead, however, Title VII has ineffectively dealt with the glass ceiling problem. That is, Title VII has been less successful at helping nontraditional workers move upward to mid-level and upper level management positions within organizations. [FN13] Most occupations remain either gender- or minority- segregated, or gender- or minority- stratified. [FN14] In elite professions in particular, women and minorities cluster at the lowest level. [FN15] The reasons why Title VII has failed to facilitate upward movement of nontraditional workers are the focus of this Article.

 
We argue that a major factor explaining Title VII's failure to deal with the glass ceiling problem is that advancement and promotion decisions occur at a different market level than entry level employment decisions. While entry level employment decisions operate for the most part in the external labor market, advancement and promotion decisions tend to take place in the internal labor market. We argue that Title VII, as interpreted, does not provide a framework capable of dealing with the problems nontraditional workers face within the context of the internal labor market. The courts' interpretation of Title VII has not been sufficiently sensitive to the subtle ways in which women and minorities come to be excluded from mid- level and upper level positions within organizations--ways so subtle that employers themselves are not always aware of them. As a result, nontraditional workers have entered internal labor markets with expectations of advancement, but when that advancement has not been forthcoming, Title VII has not been useful in enforcing the internal labor market promise.
 

We first discuss how the internal labor market operates, noting specific problems that can arise for nontraditional workers. [FN16] We then argue that Title VII does not facilitate upward movement of nontraditional workers in internal labor markets. At the same time, because Title VII has allowed greater influx of nontraditional workers from the external market into entry level positions--in fact, has required it--it is now necessary for frustrated nontraditional workers to avoid internal labor market employers and to move into jobs and career opportunities that rely on the external labor market. [FN17]
 

FNa. Assistant Professor, Department of Management, Texas A&M University, B.A., Indiana University, 1974; M.B.A., Indiana University, 1977; M.A., Indiana University, 1978; D.B.A., Indiana University, 1979; J.D., University of Nebraska, 1990.

DALLAS - A Supreme Court decision yesterday that won't allow the Dallas Fire Department to revive an affirmative-action program has some minority firefighters worried about fairness in the department.

"The Fire Department still does need the affirmative-action plan," said James Hill, a member of the Dallas Black Fire Fighters Association. "You have to have a tool in place to make people continue to do the right thing."

The Supreme Court justices, over two dissenting votes, let stand a ruling that struck down the program as discriminatory against white men.

Justices Stephen G. Breyer and Ruth Bader Ginsburg, both appointed to the court by President Clinton, voted to hear arguments in the case, to resolve apparently conflicting rulings in federal appeals courts. The votes of four of the court's nine members are needed to grant such review.

Darren Hutchinson, a law professor at Southern Methodist University, noted that yesterday's action sets no precedent and is not an indication that the Supreme Court has little patience for affirmative-action programs.

"There's just not enough evidence in the records in this case to justify using affirmative action," Hutchinson said. "I wouldn't give too much weight to this, because it doesn't set a standard."

The Dallas City Council voted in 1992 to extend a plan that called for promoting qualified blacks, Hispanics and women ahead of white men who had scored higher on Fire Department promotion tests.

Between 1991 and 1995, the Dallas Fire Fighters Association and individual firefighters filed four lawsuits challenging the promotion program.

"We felt that the program was in violation of federal law," said Mike Buehaler, first vice president of the Dallas Fire Fighters Association.

A federal judge invalidated the promotion program, and his decision finding it unconstitutional was upheld by the 5th U.S. Circuit Court of Appeals in August. The appeals court ruled that the city had failed to show the necessary "compelling interest" for such a program to promote blacks and Hispanics.

City Councilman Don Hicks said the ruling comes as no surprise.

"It's hard to start using race as a remedy when you have the political clout," said Hicks, referring to the council's diverse makeup of four African-Americans, a black mayor, and two Hispanics.

"It's the council's job to appoint a city manager who seeks diversity. The city manager hires the fire chief."

Guzman Natal, president of the Dallas Hispanic Fire Fighters Association, said there is definitely a problem with the department's demographic makeup.

"Dallas is around 30 percent Hispanic, and there are only about 120 Hispanic firefighters," Natal said.

The Dallas Fire Department is made up of 1,492 firefighters and 74 fire prevention officers. About 300 blacks and about 120 Hispanics are in the department.

In other action

The Supreme Court took the following actions yesterday:

Let the federal government continue giving companies owned by poorer people extra help in winning contracts on federal highway projects. The court turned away arguments in a Utah case that the affirmative action program adopted by Congress unlawfully discriminates against companies owned by white men.

AUSTIN - The forces behind Houston's unsuccessful 1997 Proposition A campaign are taking their crusade to the state Capitol, promoting an ambitious legislative agenda to end race and gender classifications in state government.

The Campaign for a Color-Blind America, led by anti-affirmative-action provocateur Edward Blum, launches its Capitol blitz today.

"Because the movement to end discrimination and preferences seems to be sweeping the entire country, it only seems right to me that our state representatives should also have the opportunity to vote on ending these practices by the state itself," Blum said.

Blum's group has lined up sponsors for a roster of legislation and also is organizing appearances by Ward Connerly, the University of California regent and national anti-affirmative-action figurehead.

But with the 76th Legislature already at work - notably on the methodical advancement of Gov. George W. Bush's agenda, even the sponsors of Blum's legislation are predicting the issue will receive scant attention this year.

"I would be surprised if we get a hearing," said state Rep. Robert Talton, R-Pasadena, who is filing a bill to end race and gender preferences in state hiring, contracting and university admissions. "But we have to start the debate somewhere."

State Rep. Ron Wilson, D-Houston, scoffed when he heard the Proposition A group was heading to Austin.

"I think the forces of evil are descending upon us," Wilson said. "This is a free country and they can come and espouse anything they want. . . . But they couldn't get their agenda passed in Houston, and I think they are misreading the Legislature if they think they can get it done here."

Undaunted by skeptics, Blum said he has high hopes for Talton's bill, and a separate measure that would remove from state employment forms the boxes applicants check denoting their ethnicity.

Discussions also are under way about a possible state constitutional amendment similar to those passed in California and Washington, aimed at ending affirmative action programs in state government and university admissions.

State Rep. John Culberson, R-Houston, also is filing a bill that would prevent home-rule cities such as Houston from rewording ballot language submitted by petition - a measure that harks back to a pivotal episode in the 1997 Proposition A campaign.

"What I am proposing is a truth in voting proposition aimed at ending slippery and confusing ballot language and ensuring honesty and integrity in the process," he said.

Currently mired in the courts is a legal challenge by Blum's group against the city of Houston for rewriting the ballot language for Proposition A, a gambit that both sides agree changed the outcome of the election.

As he did in Houston, Blum's legislative effort is based on the assertion that ethnic and gender classifications are regressive, discriminatory and divisive.

Although the push against affirmative action is largely a conservative one, Bush has not made the issue a priority, saying he opposes "quotas," - which are illegal - and instead favors goals for minority access to public dollars.

The state's affirmative action contracting programs are not unlike those operated by the city of Houston. Since 1991, Texas has certified so-called "historically underutilized businesses," or HUBs, which are owned by minorities and women.

The state's policy directs state agencies to make a good-faith effort to include such businesses in contracting, and graduates firms from the program when they maintain certain gross receipts or employment levels for four consecutive years.

The state's program, in keeping with court decisions on the issue, does not permit "set-asides" or specific percentages for HUB participation in contracting.

State hiring issues are slightly thornier, however. A 1997 study by the Texas Commission on Human Rights found that historically, state government did not hire minorities and women in proportion to their numbers in the labor force.

As part of the appropriations process, state agencies and universities were instructed by the Legislature to incorporate efforts to diversify the workforce.

Such efforts, however, have given rise to a series of reverse-discrimination lawsuits by Anglos against state agencies. In May, Blum's group is expected to launch a similar case of its own against the Texas Alcoholic Beverage Commission.

Also in opposition to the state's affirmative action efforts is the 5th U.S. Circuit Court of Appeals' Hopwood decision, which ended the use of race and gender preferences in university and college admissions.

Blum said his group is including higher education in its legislative program as a sort of insurance policy, to codify the current law pending appeals of the Hopwood case.

Although Blum repeatedly has turned to the courts in the past when the legislative process failed to satisfy his aims, he said recently his preference for deciding policy is with representative government.

Former Lt. Gov. Bob Bullock in late 1997 appointed a special Senate interim study group, chaired by former Mayor Bob Lanier, to review state affirmative action programs and make recommendations.

The committee's final report, which relied on testimony, legal rulings and various studies, includes a series of modest recommendations aimed at continuing affirmative action policies at the state level.

In a letter to the committee, Lanier noted that affirmative action continues to play an important role, "until such time as we have a level playing field."

Lanier, who as a term-limited mayor waged a brutal campaign against Blum to maintain the city's affirmative action program, concluded similar policies are still needed at the state level.

"I think it would be wrong to go in either direction, in terms of making any reductions or a major increase," Lanier said. "It would come at too great a political cost and I think it would be unwise, divisive and it could really hang up the Legislature."

In the area of higher education, the panel calls for increased funding for scholarships, student outreach programs and work-study efforts. It also suggested development of an alternative admissions criteria that targets economically disadvantaged students.

The report also suggests creating a mechanism to allow community college students with high grade point averages to transfer to four-year universities.

In the area of state contracting, the panel recommends few changes other than general improvements to the current system, such as bonding and technical assistance to small or disadvantaged businesses.

Similarly, the bipartisan panel recommends state hiring practices maintain the current policies for minority recruitment.

Lanier said evidence collected by the committee showed significant race and gender disparities in the upper management levels at state agencies.

"There was not much discrimination at the entry level, but once you move up the ladder there were not a lot of women and minorities represented," Lanier said. "That is one area where I still think affirmative action has a strong use."

State Rep. Garnet Coleman, D-Houston, noted that Texas' minority population during the next 10 years is expected to exceed that of Anglos.

"I do not believe we live in a colorblind society, and I also don't believe the concept of historic discrimination has outlived its usefulness," Coleman said. "We have to be mindful of what is good for Texas, and to bring a point of view that is exclusive as opposed to inclusive is the wrong point of view."

Employment discrimination is still a big part of how America does business, says Barbara Reskin, author of The Realities of Affirmative Action in Employment. Affirmative action is needed to counter entrenched discriminatory hiring practices that cause inequality and keep the need for affirmative action alive.

The report, published in August 1998, demonstrates how age-old recruitment practices such as informal networks and employee referrals result in discrimination.

"A common misconception is that race and sex discrimination no longer exist," said Reskin. Citing a compelling bevy of studies and statistics Reskin demonstrates discrimination in employment. "Discrimination in job assignment and promotion - whether the result of intentional acts or customary business practices - is still common."

The book shows that thirty percent of white workers held managerial or professional jobs in 1996, while only one in five African Americans and one in seven Hispanic workers did. Moreover, 58 percent of white women, 58 percent of Hispanic women, and 60 percent of African-American Women would have to hold different occupations to be distributed across occupations in the same way as White men.

"Discrimination is not simply the result of deliberate attempts to discriminate," said Reskin. "While some discrimination results from employer's active preference for one group over another, much of it stems from simply by doing 'business as usual'."

Word of mouth recruitment - identifying job candidates though referrals by current employees -- is the most commonly used recruitment method because it is cheap, the report finds. But employee referrals maintain the race, ethnic, and sex composition of a firm's workforce because people's acquaintances tend to be of their same sex and race.

"When employers fill jobs though informal networks, minorities and women don't learn about jobs for which they are qualified," said Reskin, "even when those jobs are abundant and even in firms in their own neighborhoods."

Because they are often unintentional, discriminatory practices can be hard to detect and eliminate. Common discriminatory but well-intended hiring practices include requiring credentials, skills or experience beyond the requirements of the job, using racial or gender stereotypes when lacking complete information about applicants, and designing jobs based on the assumption that one sex or the other will hold them.

The Realities of Affirmative Action in Employment is a compendium of sociological research into the causes and consequences of workplace discrimination. Affirmative action programs must be strengthened, the report concludes. Current trends towards dismantling affirmative action, such as proposition 206 in California, are "a costly and dangerous experiment that the United States cannot afford," said Reskin.

Reskin finds that much of the rising anti-affirmative action sentiment is based on fallacies and misconceptions: mainly three central myths: that discrimination no longer exists, that affirmative action reduces opportunities for whites, and affirmative action undermines merit based hiring.

"Exacerbating the confusion is the fact that affirmative action has become a political football," said Reskin. "The object of summarizing the realities of affirmative action is to provide a foundation for informed public discourse and social policy."

The ASA commissioned the study to help bring sociologists' special expertise on the topic more accessible, said Felice Levine, ASA's Executive Officer.

"Affirmative action has been politicized and hotly debated, but it has also been rigorously researched by sociologists," said Levine. "This report brings that research to a wide audience. We hope it will foster intelligent public debate."

Barbara Reskin is Professor of Sociology at Harvard University and the co-author of several books and articles on job segregation and discrimination in the workplace. Men and Women at Work: Sociology for a New Century (Sage, 1994); Job Queues, Gender Queues: Explaining Women’s Inroads into Male Occupations (Temple University Press, 1990) and Sex Segregation in the Workplace: Trends, Explanations, Remedies (National Academy Press, 1984 ).

The American Sociological Association, founded in 1905, is a non-profit membership organization dedicated to advancing sociology as a scientific discipline and profession serving the public good.