The American State and Social Engineering: Policy Instruments in Affirmative Action
DESMOND KING*
  *Nuffield College, University of Oxford 
Copyright © 2007 The Author; Journal compilation © 2007 Blackwell Publishing
ABSTRACT
       

The American state is conventionally depicted as inactive and organizationally weak compared with the state in comparable industrial democracies, and it is sometimes depicted as weak compared with the private sector's capacities to effect change. This interpretation stems from applying an inappropriate Weberian model of stateness. This article examines the way in which measures to implement affirmative have been employed through the policy instruments of quotas to reengineer the divisions between key groups in American society. Placed in historical context, affirmative action illustrates a powerful activism associated with the American state conceived as an institution engaged in setting and monitoring national standards.
  
  
Article Text
The American or U.S. state is conventionally depicted as inactive and organizationally weak compared with the state in other industrial democracies, and sometimes depicted as weak compared with the private sector's capacities to effect change (Dobbin and Sutton 1998; Evans, Rueschemeyer, and Skocpol 1985). On Peter Nettl's (1968) famous scale of "stateness," the United States was by far the least state-like among Western democracies, in terms of Weberian criteria of bureaucratic development, administrative resources, and public sector capacities. But this Weberian framework is too limiting for the U.S. case. It masks or overlooks a good deal of policy intervention and direction originating in the late nineteenth century and escalating through the years of the New Deal, war mobilization, and the Great Society. In this article I examine the way in which measures to implement affirmative action or positive discrimination have been employed through the policy instrument of quotas to reengineer the divisions between key groups in American society, groups based on race and ethnicity. Placed in appropriate historical context, affirmative action illustrates a powerful activism associated with the American state conceived as an institution which engages in the setting and monitoring of standards. As a policy instrument of the federal government, the system of quotas linked to affirmative action was formulated in the 1960s. It has been implemented since then in a federal policy which sets comprehensive targets and specifications affecting both the public and private sectors (Drake and Holsworth 1996; Sabbagh 2003, 2004; Skrentny 1996, 2001, 2002).

Historically, the term "state" is considered problematic when applied to the United States (except perhaps in foreign policy; Krasner 1978). The horizontal separation of powers among the executive, legislature, and judiciary often precludes the identification of a national source of authority in the way commonplace to centralized states, a tendency strengthened by federalism. The dominant emphasis in the polity's political culture is that of being a strong nation with a weak state, that is, a political system less centralized, less interventionist, and less bureaucratic than that found in comparable advanced democracies, including those with strong federal systems such as Australia or Germany. But reluctance to use the term "state" has gradually dissipated as major scholarly works by social scientists and historians have exposed the limits of perpetuating the presumption of American exceptionalism in respect to its state (Brinkley 1995; Karl 1983; Rung 2002; Skocpol 1992; Skowronek 1982; Stears 2002; Weir and Skocpol 1985). Disagreement about nomenclature cannot mask the growth of the most powerful nation in the second half of the twentieth century, which is expressed in commonly identifiable indicators of statehood (Gonzalez and King 2004; Nettl 1968) from the highly measurable, such as level of expenditure (Kettl 2002; Light 1999) and military capacity, to the highly symbolic, such as myths about being "one people" (Gerstle 2001; Stuckey 2004).

What makes affirmative action especially valuable for a discussion of policy instruments, to illustrate the activism and intervention of the U.S. state, is the scale of policy change its proponents envisage, despite often modest expectations among its early champions. Affirmative action does not involve simply either a uniform regulation across the United States (for instance, that no guns may be carried within a certain distance of schools) or a behavior-driven measure (for instance, raising the drinking age to 21 in all states to tackle drunken teenage driving). Rather, the policy of affirmative action mixes several aims including bolstering barriers against racial and gender discrimination, fostering new practices in the workplace and in public institutions such as schools, creating social relations of a different character for America's future by modifying how groups perceive each other, and denuding historical injustices of their political salience.

This range of ambitions makes it more difficult, although not impossible, to measure the success or effectiveness of affirmative action. Certain indicators—such as the number of women or members of groups previously discriminated against in higher education or senior management or senior federal administration (Aberbach and Rockman 2000)—can be constructed but these will not measure the more nebulous aspects of affirmative action such as changed attitudes and future societal trajectory shaped by its terms. To achieve purchase on this complexity, I focus on a single policy instrument, quotas.

In order to better understand the scope of affirmative action, it can be usefully compared with another systematic incident of social engineering, that of eugenic breeding, the main instrument of which was sterilization, a practice supported by the federal government (Trent 1994) and enacted by state governments between the 1920s and 1960s, and legitimated in the U.S. Supreme Court judgment in the case Buck v. Bell (274 US 200 1927). Although both affirmative action and eugenic breeding had similar ambitions to effect major transformations in fundamental aspects of the American social order, they had one major difference. The eugenic project presumed the validity of differentiating hierarchically between groups in terms of race and ethnicity, and sought to entrench that vertical hierarchy through a policy of sterilizing the unfit and inferior. Advocates of affirmative action reverse this assumption seeking to build a nonhierarchical order in which groups enjoy horizontal equality and in fact will eventually wither away as policies of preferential treatment cancel historical fetters upon equality of individual opportunity and outcome.

"Policy instrument" means tool of government or state policy. It assumes relevance as an expression of policy instrumentation. Politicians and policymakers identify policy instruments as means to effecting ends (Hood 1983). Lascoumes and Le Galès (2007) define a public policy instrument as "a device that is both technical and social, that organizes specific social relations between the state and those it is addressed to, according to the representations and meanings it carries." In respect to affirmative action, this end is a daunting one—a social engineering of society to redress enduring historical injustices and to establish barriers against future harms. Such efforts are intended to redraw the boundaries of membership of the polity to strengthen the bases of citizenship, particularly for those Americans who find themselves distinguished by others in terms of race, ethnicity, or national background (see King 2005). The policy of eugenic breeding was also designed to redraw boundaries but in a hierarchical way (Gallagher 1999).

The article is structured as follows. In the next section, I identify six factors which help social engineering exercises to gain momentum and succeed politically. Comparing affirmative action with eugenic breeding is a useful way of demonstrating how these factors operate temporally, although this comparison is not intended to imply any constancy in the causal factors. I then place the policy of affirmative action in the context of a group-bounded polity in which relations between groups as members of a single society are renegotiated over time. This discussion is a prelude to a detailed overview of the development of the key policy instrument for affirmative action—quotas—in the major areas of employment, university admissions, and business ownership. I then conclude the article.

  
  
Eugenic Breeding and Affirmative Action
       

What factors justify social engineering programs? Why do they become part of the political agenda? To consider the circumstances under which social engineering projects garner support, it is useful to think about the eugenic breeding and affirmative action initiatives in terms of six factors. To some extent, this concatenation of factors creates a critical juncture in policymaking when a particular course of public policy is set and its realization locked into a selected policy instrument.

First, we can attend to the nature of the perceived problem addressed by the social engineering exercise. For eugenicists, it was the corruption and dilution of the American "race" by inferior immigrants and the so-called "feebleminded" parents which excited their concern (Black 2003; Grekul, Krahn, and Odynak 2004; Kevles 1986). For affirmative action advocates, it is the persistence and depth of inequalities rooted in the legacies of the United States's history, specifically not only its mistreatment of African Americans but also other groups such as American Indians and many Latinos and, from 1967, women: without action to remedy these entrenched inequities they would endure and reproduce. Thus, each exercise in social engineering posited a clearly defined problem and stressed the deleterious implications for American society if the problem remained unaddressed and was allowed to fester. Thus, social engineering projects are associated with definable social philosophies about the social order: egalitarianism in affirmative action, hierarchy in eugenic breeding.

Second, once the problem is defined, the role of experts and of expert knowledge in defining the problem and solutions is often crucial in determining the way in which policy instruments are developed. This is an important contrast between eugenic breeding and affirmative action. In respect to eugenic sterilization, experts campaigned long and hard marshalling data in support of their cause. Expertise in respect to affirmative action was generally more tentative: there was great conviction about the seriousness and urgency of a problem—entrenched historical patterns of inequality—which had to be addressed, and politicians from Lyndon Johnson through Bill Clinton and former Secretary of State Colin Powell have underlined its significance. But settling upon a course of action to deal with this was more tentative; however, once quotas were agreed upon as a policy instrument, they were defended by experts, particularly lawyers and human resource managers.

Third, how politicians and policy makers respond to the varying political opportunities to pursue or challenge social engineering projects is obviously decisive to their success or failure. Eugenicists enjoyed many advocates and supporters in government as have those advancing affirmative action measures (King and Hansen 1999). Crucially, both initiatives benefited from cross-partisan support at key moments, especially in their early stages. Furthermore, in both cases, the U.S. Supreme Court provided endorsement at key moments in the adoption of policy, legitimating the policy but specifically providing the measures chosen as policy instruments to achieve them about which some Americans had real doubts. The Court has a singular constitutional authority in the U.S. polity.

In its decision on Buck v. Bell (274 US 200 1927), the U.S. Supreme Court upheld a state law from Virginia permitting the use of sterilization to prevent designated persons from reproducing. The Court famously declared that "three generations of imbeciles" were sufficient. This ruling, which remained in place until the 1970s, enabled 27 states to adopt similar sterilization laws (and close to 60,000 Americans were sterilized) (Black 2003). In respect to affirmative action, the U.S. Supreme Court issued decisions upholding the implications of the 1964 Civil Rights Act, particularly Title VII banning job discrimination. In Griggs v. Duke Power Co. (401 US 424 1971), the Court ruled against the use of intelligence tests in making decisions about whom to hire because even if they were neutral, such tests often operated to "freeze" the status quo of discriminatory practices. As a consequence of this judgment, employers had to demonstrate statistical parity between the racial makeup of their employees and the local population from which the workforce was drawn. In Fullilove v. Klutznick (448 US 448 1980), the Court held "set asides"—that is, setting a percentage of jobs directly eligible for minority candidates only—to be constitutional.

This minority business enterprise ruling made clear that the federal government had greater authority than either state or local governments to require the adoption of instruments of affirmative action in certain circumstances. It demonstrates the active U.S. state, and as is well known, there have been significant judicial rulings about the constitutionality of quotas in admission to universities with the most recent (2003) Court decision permitting race to be a factor in university administration's decision making about allocating university places although now finding explicit race quotas unconstitutional.

Thus, the Court has been a key actor in the evolution and implementation of affirmative action just as it was for eugenic breeding. It constituted a political opportunity structure largely favoring social engineering and its constitutional authority, once invoked in a ruling, is not easy to overturn.

A fourth factor in each case is the nature of the target population and in particular its political strength or weakness. Here, there is a fundamental difference between eugenic breeding and affirmative action experiments in social engineering. Eugenicists targeted the marginal and the weak in society, determined to excise their reproductive powers in the pursuit of a racially pure and stronger national stock. In contrast, affirmative action measures are designed to enhance the presence and capacity of their target population's participation in American society and in direct reverse to eugenics, to make their future participation much greater. But like victims of eugenics, the target population for affirmative action, by definition, occupies politically weak positions in the polity.

Fifth, the practicality and validity of the policy instruments available to advocates of these two efforts in social engineering are important. Here, there are similarities in that advocates of each exercise promoted a set of carefully specified and easily comprehended policy instruments (and in the case of eugenicists, claimed scientific validity for their deployment). Eugenicists singled out sterilization as a key policy instrument to effect their aim of a better demographic future as they saw it. Affirmative action advocates have found in the measures of preferential employment expressed in quota places in public institutions such as universities, public sector agencies, and federal contract holding private employers an equally strong policy instrument for realizing their ends. Counting and measuring are crucial to both instruments because they permit monitoring and evaluations of success.

Last, the political power or weakness of opposition—or even the existence of opposition—is crucial to achieving social reform on the scale of these two programs. In both cases, critics were wrong-footed and politically weak at key stages in the policy's development. There were critics of the scientific validity of eugenic arguments for sterilization as an effective tool of "race" reproduction and critics of the basic assumptions guiding their scientific analysis. The fundamental concerns of these critics were vindicated in due course but not before eugenicists had persuaded federal and state lawmakers to implement schemes based on their flawed scientific arguments and evidence.

Critics of the use of quotas as a means to pursue affirmative action voiced concerns in the 1960s, but were muted by a political atmosphere in which such questioning quickly provoked charges of racism and discrimination. In the 1970s, it was the Republican administration of Richard Nixon which oversaw a decisive push in quota-based policy to realize his Democrat predecessor's commitment to affirmative action. The basic position of critics has endured—that any system of preferential treatment commits new injustices against worthy competitors and fails to deliver the promised amelioration—and has gained support with the expansion of affirmative action quotas. The African American scholar Glenn C. Loury (2004) captures a common complaint about policy instruments such as quotas: "Affirmative action has a paradox: blacks seek equality with whites, but by accepting special treatment, they call attention to their own limitations. Meanwhile, by the very act of granting black demands, whites exercise a noblesse oblige available only to the powerful. While equality is the goal, this manifestly is not and never can be an exchange among equals."

In respect to each of these six factors, considerable differences distinguish the examples of eugenic breeding and affirmative action. Eugenicists marshalled extravagant claims about the ability of measures such as sterilization to shape future generations, and based these claims on the so-called scientific evidence. Champions of affirmative action instruments, such as quotas, were more modest about scientific evidence. Rather, it was upon the inherent importance of the case to address enduring and deep historical inequalities that their arguments rested. Furthermore, without addressing these historical legacies, they argued, the United States would simply face increasing inequality and more of the same. This was a point Lyndon Johnson underlined as president and after he left the White House in 1969: civil rights legislation could not magically conjure up a level playing field for groups in American society who had been the victims of systematic efforts to segregate, marginalize, and undereducate them. In December 1972, weeks before he died, Johnson spoke to a civil rights conference in Austin, Texas and accepted that the "black problem today" was not a local problem but one confronting "this whole nation." Progress required overcoming this "unequal history," an experience applicable to other groups in American society too (quoted in Dallek 1998, 621–622).

But there are some points of similarity between eugenic breeding and affirmative action initiatives despite their supporters' commitment to obverse conceptions of the social order. Both programs benefited from the horizontal separation of powers in the U.S. polity, with the judicial system providing decisive support for their policy instruments at key moments. Furthermore, each program benefited from clearly formulated policy instruments—eugenic sterilization and preferential treatment, respectively—which could be measured and assessed over time. These similarities alert us as to how useful social engineering programs are projects in which policy instruments become paramount in the way analyzed by Lascoumes and Le Galès (2007) when they write, "Public policy instrumentation is … a means of orienting relations between political society (via the administrative executive) and civil society (via its administered subjects), through intermediaries in the form of devices that mix technical components (measuring, calculating, the rule of law, procedure) and social components (representation, symbol)." Because social engineering projects envisage very substantial change in the relationships between existing groups, a clear policy instrument is vital to the cogency of their advocates' case. This point can be seen more fully by appreciating how social engineering seeks to restructure group-based boundaries of political membership.

  
  
Social Engineering and the Boundaries of Membership
       

Both eugenicists and advocates of affirmative action want state policy employed to redefine the position of some members of society. This task occurs within historical contexts shaped by the United States's group-based ideology of nationhood (King and Smith 2005; Smith 1997). It is the boundaries of these group divisions which prompt social engineering. This can be illustrated in respect to the historical pressures driving such initiatives as affirmative action.

Challenges to revise the boundaries of internal membership, to enrich senses of belonging, and thereby to renew the American nation take several forms (Balfour 2003; Hattam 2005). Proponents of change are motivated by the ways in which the language and practice of group distinctions, often associated with inequalities, shape nation building in America.

First, there are revisions to give or withhold some groups the rights of citizenship. The wartime treatment of Japanese Americans illustrates the latter while African Americans' fight for civil rights is an instance of the former. Enlarging inclusion is the most familiar type of revision and dominates the conventional narrative of America's shift to democratic inclusiveness. But revision can also be exclusionary as some Arab Americans have recently found and as judicial rulings about legal aliens since the Patriot Act of 2001 and the President's Executive Order (EO) issued on November 13, 2001 permitting detention and military trials for noncitizens accused of terrorism attest (on EOs see Mayer 2001; Robinson 2001). (Eugenicists were plainly committed to an exclusionary trajectory.)

Second, boundary revision can involve some specific redress or remedy or compensation for an identifiable violation of legal rights or for a historic injustice. The success of Native American claims about treaty violations is a case of such revision as is the payment of compensation to victims of sterilization by American states.

Third, boundary revision can be undertaken to anticipate and preclude future inequalities or violations. Measures taken under this ambitious strategy can also themselves become sources of future conflicts about group membership and the parameters of democratic inclusion. Affirmative action policies are a case in point: their rationale is to improve the future prospects of those eligible for these benefits but their implementation stokes existing lines of group cleavage and for some Americans forge new tensions (Davis 2002; Huntington 2004). Minority–majority electoral districts are equally fraught (Canon 1999).

Common to each type of boundary revision is a shared assumption: group hierarchies are an inadequate basis for a genuinely inclusive ideology of nationhood given America's community of groups. Yet, Americans' and the United States's beliefs about group hierarchies have a deep hold in the nation: just as the Supreme Court upheld color blindness so it accepted color-conscious policies to address the effects of failure to achieve the former. Indeed, the endurance of group distinctions in itself encourages the retention of hierarchical assumptions, usually with fixed views about the unsuitability of some groups to be full members. Compensation and reconciliation measures are designed to deepen the affected individuals' and groups' sense of belonging and membership in the "one people" nation by expunging and recognizing the damage of earlier episodes in nation building. To move forward as "one people" requires remembering how group diversity and nation building has been inadequately reconciled in the past.

Understanding the demands and effects of boundary revisions is often associated with the Commencement speech delivered by President Lyndon Johnson at Howard University in 1965 when the university bestowed an honorary degree on the president. Johnson made justice for African Americans a federal priority, as a response to the historic injustice experienced by these citizens. The president declared, "in far too many ways American Negroes have been another nation, deprived of freedom, crippled by hatred, the doors of opportunity close to hope." Signalling great optimism, Johnson concluded that "it is the glorious opportunity of this generation to end the one huge wrong of the American Nation and, in so doing, to find America for ourselves, with the same immense thrill of discovery which griped those who first began to realize that here, at last, was a home for freedom" (Johnson 1965).

But the implications of Johnson's address were wider than simply setting an agenda for African Americans. Unintendedly, the speech cast a light on many aspects of the nation building process and of how membership and international influences, especially as a consequence of America's standing as a model of liberal democracy, mediated the content of nationhood; and the speech conveyed, although perhaps in ways the President failed to appreciate, just how far boundaries of membership would need to be revised to make democratic inclusion of equal importance to assimilation in American nationalism. Thus, although his speech addressed the circumstances of African Americans only, it had implications for Native Americans, Mexican Americans, and other groups historically distinguished by race, ethnicity, or national background.

  
  
Quotas and the Development of Affirmative Action
       

The principle of affirmative action posed a formidable challenge to its advocates. What policy instrument could possibly compensate for decades of historical injustices and discrimination against African Americans and other minorities in the United States?

A taste of the problem emerged during the Second World War when President Franklin D. Roosevelt established a fair employment practices committee (FEPC) and issued an EO banning job discrimination on the basis of race. These measures were responses to a threatened mass mobilization by African Americans in Washington at a time when their labor was vital to war mobilization (Kryder 2000). The FEPC proved to be a forum whose proceedings unmasked profound and systematic discrimination against African Americans in most public and private sector employment sources. Its members heard complaints from discriminated workers and issued injunctions to employers to desist discriminatory behavior; it was temporary and its instructions were often ignored.

The issues confronted by the FEPC did not evaporate; in fact they intensified as the civil rights movement mobilized from the 1940s forcing judicial overturning of segregation (in the 1954 Brown case) and eventually passage of the Civil Rights Act of 1964 and the Voting Rights Act of 1965. The U.S. federal government was much exercised about employment equality within its own ranks and wrestled with the question of how to respond in a way which tackled enduring historical injustices. The dilemma, ultimately resolved in affirmative action and preferential treatment, was anticipated by the Executive Secretary of the President's Committee on Government Employment Policy as he contemplated, in 1961, how best to redress the massive segregation imposed upon African Americans in the civil service. "A mere head count showing the absence of negro personnel simply is not enough information upon which to make a judgment, for the reason may be, and in many situations is, the absence of Negroes on certificates which the agency is using." He recognized the mounting pressure for some sort of affirmative action or preferential treatment scheme.

Our critics would then say that the hiring practices of the agencies should be reviewed. But to do this either we or the agencies must be able to identify, through a review of certificates, those Negroes who have been considered and determine whether or not they have been passed over. With the absence of any identifying racial information on the certificates, it becomes impossible to review past practices. The only possibility lies in a current review on a day-by-day basis, based almost entirely on identifying those Negroes who appear for personal interviews. Even if such identifications are made (and it may be possible to do so) there remains the further problem of determining the relative significance of those Negroes passed over as compared with the whites who were hired before any finding of discrimination can be made. On any large scale this would be a very involved process.

To correct such a pattern required a different course: "the only alternative to this difficult process that I can see at present is simply to take the position that where there are no Negroes there ought to be some, and direct the agencies to find and employ them." But this strategy was also problematic: "This comes close to the 'quota' idea, and in any event would put the program in the position of deliberate preference for Negroes" (quoted in King 1995, 209 citing the President's Commission on Government Employment Policy records in February 1961).

This exposition anticipated many of the problems in ensuring equal opportunity for African Americans in the federal government. There was evident anxiety about using a racially based quota hiring system as the principal policy instrument for achieving equity of employment but no obvious alternative presented itself to monitor and demonstrate enforcement.

Such a quota system, however, or some policy instrument like it, was not an automatic implication of the 1964 Civil Rights Act. Title VII of the 1964 Act disallows any discrimination, by any public agencies, private employers (with more than 15 employees), and labor unions, on the basis of race, color, religion, national origin, and sex. Instead, quotas became a logical policy instrument as a consequence of President Johnson's initiatives to achieve affirmative action. Thus, EO 11246 issued by President Lyndon Johnson on September 24, 1965 set out the principle of affirmative action as a goal to be realized subsequent to the passage of civil rights legislation.

The term "affirmative action" had been formulated by a young African American lawyer in Texas, Hobart Taylor Jr., at then Vice President Johnson's request at the time of EO 10925 establishing the President's Committee on Equal Employment Opportunity (in 1961). Taylor recollected his formulation: "I was searching for something that would give a sense of positiveness of performance under that Executive Order and I was torn between the words 'positive action' and the words 'affirmative action.' And I took 'affirmative' because it was alliterative" (quoted in Lemann 1999, 162).

EO 11246 set out the prime policy instrument to address historical injustices toward African Americans and other minorities, and rather than simply confirming the antidiscrimination measures in the Civil Rights Act, the Order required active responses by private firms receiving federal grants to address patterns of hiring discrimination (monitored by the Office of Federal Contract Compliance in the Department of Labor). However, the specification of this policy instrument was quite general in its early formulation. Thus, among its measures, EO 11246 "enjoined government contractors and others receiving government funds to promote hiring of blacks and other minorities." At the time, this instruction presented a relatively low-cost solution to the challenge of integrating African Americans into society, not just in terms of equal civil rights, but also in terms of overcoming economic inequalities.

But EO 11246 also pointed policy toward a quantification approach because to be in receipt of federal funds through contracts, employers needed to demonstrate that their hiring practices complied with the requirement to hire African Americans. It is difficult to imagine how this compliance could be assessed without some resort to the compilation of statistical data about the racial (and ethnic and gender) profile of a given firm's workforce, especially relative to the composition of the local workforce, and as we will see the U.S. Supreme Court, the key interpreter of policy instrumentation in affirmative action, also came to rely on a quota-based understanding of this agenda. Such an approach was anticipated in the remarks of the 1961 President's Committee cited previously.

This quantification bias has two immediate effects. First, it imbues racial and other demarking categories with policy significance seemingly at variance with the color-blind and antidiscrimination thrust of the Civil Rights Act of 1964. Second, it makes quantifiable and quantified data a central indicator of policy success and therefore a part of social reality (Desrosieres 1998). Such data are the lifeblood of administration in Western states. In any official assessment of a policy such as affirmative action, such data, including its updating and augmentation, plays a central role.

EO 11246 applied to other minorities too and, from 1967, to women. Expanded decisively by the Nixon administration (through his support of the trade union Philadelphia Plan for hiring minority workers as apprentices on federal contract work), the major policy instrument of affirmative action quickly became specifying targets or quotas in a range of areas which had to be satisfied. Thus, holders of federal contracts had to employ a certain percentage of African Americans and other minorities; the federal government's and state governments' workforce had to be made representative of America's minorities; and admission to universities, particularly state universities, was recast to provide preferential access to members of ethnic minorities even when such preferences admitted candidates with lower test scores than some white applicants. This last measure has reverberated for critics of affirmative action and is a policy they have struggled to dismantle.

Quotas to achieve affirmative action have thus been implemented in several areas of American life. I deal in detail with that of education below but preliminarily it is important to underline the importance of the U.S. Supreme Court in facilitating this diffusion of affirmative action. Although Title VII of the 1964 Civil Rights Act enacted a ban on job discrimination based on race, sex, or national origin, in fact the program of affirmative action required combining this color-blind principle with a quota-based system premised on recognizing race or other differences (Lieberman 2002). This paradox is part of American history and the Court has been a key agent accepting the constitutionality of this paradox. We have already noted the Court's decisions in Griggs v. Duke Power Co. (1971) and Fullilove v. Klutznick (1980) permitting set aside quotas. In Metro Broadcasting Inc. v. Federal Communications Commission (FCC) (497 US 547 1990), the Court upheld, by 5:4, the right of Congress to enact policies that favored minorities; in this case the ruling concerned ownership of broadcast licenses. The FCC wanted to extend licenses and thereby increase diversity of ownership among African American, Hispanic surnamed, American Eskimo, Aluet, American Indian, and Asian American groups.

But since the 1980s, the Court has adopted a more restrictive approach to the use of quotas as a policy instrument in contract enforcement. Thus, in Richmond v. J. A. Croson Co (488 US 469 1989), the Richmond city council's set aside ordinance (the Minority Business Utilization Plan), which required white contractors who received city construction contracts to employ minority subcontractors and suppliers for at least 30% of the total dollar amount, the Court rejected the plan as unconstitutional. The Justices ruled by 6:3 that such a policy instrument conflicted with the equal protection clause of the Fourteenth Amendment. The Court established a strict scrutiny standard applicable to all such schemes defined by three criteria: skepticism, any preference for a racial group had to be examined carefully; consistency, any measure had to be consistent with the Fourteenth Amendment; and congruence, any measure should satisfy equal protection in the Fifth Amendment too.

This standard was underlined further in the Adarand Constructors Inc v. Pena (515 US 200 1995). In a 5:4 judgment, the Court held that a 1990 Transportation Department program, which awarded bonuses to highway contractors who gave at least 10% of their contracts to "disadvantaged business enterprises" failed to satisfy strict scrutiny. Indeed, two justices expressed preference to ban quotas but majority of the justices accepted that the "lingering effects of racial discrimination" in U.S. society justified some positive measures.

Adarand was a significant judgment. While the Clinton White House defended its commitment to affirmative action policy, it also instituted, in response to the Court's ruling, a review of federal programs to end any program which created quota requirements. That the Court is deeply divided about the use of affirmative action as a policy instrument is suggested in comments from one of the five majority justices in Adarand, Sandra Day O'Connor, when she wrote, "We wish to dispel the notion that strict scrutiny is strict in theory but fatal in practice." The persistence of racial discrimination means the Court has to take account of this problem in reaching its judgments in decisions about affirmative action quotas.

Despite the continuing legal wrangles about the appositeness of quotas as a policy instrument for affirmative action, the pursuit of such quotas in federally funded contracts to private employers and public agencies since the 1970s has induced a sea change in employment practices. Most private firms, and especially those with large number of employees, have instituted some sort of employment strategy to accommodate the pressures associated with affirmative action demands. Many have appointed affirmative action officers or diversity officers and set out in written form the equal employment opportunities, which they aim to satisfy in their hiring practices and promotion reviews (Dobbin and Sutton 1998; Dobbin et al. 1993). These are generally still linked to a quota-based framework—because this provides data which can be publicly defended as quantified evidence about affirmative action responsiveness—although increasingly described in terms of diversity management.

  
  
Quotas in Education
The U.S. Supreme Court has played a central role in defining the limits and scope of affirmative action in university policy. Universities are often touted as an important medium of advancement in U.S. society. Before the 1960s, they were the bastions of white Americans. This was especially true of elite universities and prestigious graduate schools in law, medicine, and business. A small percentage of African Americans had long benefited from black-only colleges and universities but only few had gained admission to elite institutions. Elite universities, public and private, recognized the impetus and implications of the laws enacted in the 1960s. Admission officers hastened to widen their access criteria and to make their student bodies representative of the groups composing American society.

The first major judicial test of this strategy came in the famous Regents of University of California v. Bakke (98 US 2733 1978) in which a white student sued the Davis Medical School for failing to admit him to its medical school despite having higher marks on the entrance test than several applicants from ethnic minorities who were accepted because the school maintained a 16% minority quota. The Court's judgment was less than pellucid. While the justices rejected as unconstitutional the use of quotas in admissions policy, it permitted universities to consider race as one of several "plus factors" in selecting applicants. Justice Lewis Powell wrote in Bakke that diversity "clearly is a constitutionally permissible goal for an institution of higher education."

In 1995, the Regents of the University of California banned the use of race in admissions. Texas adopted a similar reversal of earlier admissions policy. The changes have reduced the number of minorities attending leading state universities, especially at the graduate level in law schools. But such reversals have been the dominant trend since the early 1990s.

On March 27, 2001, a federal judge (U.S. District Judge Bernard A. Friedman) rejected the use of race as a factor in higher education admissions ruling that the University of Michigan Law School's practice of accepting minority students with lower grades and test scores than white students is not justified by the Law School's commitment to achieve racial diversity. The justice rejected the university's claim that racial diversity is a "compelling" state interest and found the Law School's admissions policy to be unconstitutional.

Friedman's judgment was comprehensive. He articulated the objections of critics of quotas: "By using race to ensure the enrollment of a certain minimum percentage of underrepresented minority students, the law school has made the current admissions policy practically indistinguishable from a quota system." He added, "All racial distinctions are inherently suspect and presumptively invalid. Whatever solution the law school elects to pursue, it must be race-neutral. The haphazard selection of certain races is a far cry from the 'close fit' between the means and the ends that the Constitution demands. If the law school may single out these racial groups for a special commitment today, there is nothing to prevent it from enlarging, reducing or shifting its list of preferred groups tomorrow without any reasoned basis of logical stopping point." He found that white applicants with high scores were failing to win admission to the Law School when African Americans were admitted with lower scores.

Friedman's decision conflicted with a decision in the same district's Court of Appeals a year earlier which permitted the University of Michigan to use racial preferences in undergraduate admissions policy on the ground of diversity. Both cases went to the U.S. Supreme Court in 2003. The justices found the use of a race quota in admission decisions unconstitutional but permitted as constitutional the use of "diversity" factors by admission tutors. In Grutter v. Bollinger (No 02-241 WL 21433492 2003), the Court allowed the use of race in creating a diverse student body at Michigan's Law School while in Gratz v. Bollinger (No 02-516 WL 21434002 2003), the Court disallowed Michigan's affirmative action admissions policy for undergraduate selection. This dual decision was much less of a firm rejection of affirmative action quotas than its critics sought.

  
  
Conclusion
       

The use of quotas in such areas as university admissions to advance affirmative action has been controversial in the United States (Mendelberg 2000; Sniderman and Piazza 1993). Indeed, for its most ardent critics, quotas are as objectionable—principally because of their dependence on racial sources of political distinction—as eugenic sterilization would be to most Americans, were its revival now proposed. I argue that that controversial status of affirmative action stems, in large part, from ignorance of the historical circumstances in which it originated, an ignorance which makes assessment of affirmative action as a policy instrument difficult.

Affirmative action is not something designed and implemented in a historical vacuum—it has been a social engineering initiative responsive to the persistence and entrenchment of historical inequalities and racist legacies. This context places affirmative action quotas at the most extensive end of the fivefold policy instruments framework developed by Lascoumes and Le Galès (2007) and gives weight to their observation that "every public policy instrument constitutes a condensed and finalized form of knowledge about social control and ways of exercising it." As a policy instrument, the idea of quotas has the appeal of being measurable and a means of setting out achievable targets. For its critics, of course, that very measurability is crude, which produces distorted outcomes harmful not only to those passed over despite qualifications, but also to those given the advantages of preferential treatment. But as Lascoumes and Le Galès note, because "it is exceptional for a policy, or even a program for action within a policy, to be noninstrumental" it is hardly surprising that affirmative action quotas have had such a dichotomous political impact.

There are two aspects of the United States's historical context which impinge especially upon the emergence of affirmative action as a policy and quotas as a policy instrument to address historical injustices. First, until the 1960s, the United States was a hierarchical group bound society (maintained through judicial rulings) in which the aspiration of equality of opportunity and treatment was just that for many Americans—a rhetorically empty aspiration. This set a historical agenda for renegotiating the boundaries between groups' membership of the American polity. Second, as Nicholas Lemann (1999) has shown, the shift to standardized testing in admission decisions by universities, formalized from the 1960s, conflicted directly with the expectations of historical compensation mobilized in the successes of the civil rights movement. This was a particular and an unforeseeable historical clash.

The former historical context—America's legacy of hierarchical divisions—explains why affirmative action quotas rapidly assumed the qualities of a major social engineering exercise rather than becoming an incremental policy innovation. The latter effect—the shift in university admissions criteria—shows how a social engineering program using policy instruments to affect the highest echelons of American society presumed a strong and activist U.S. state, often underestimated by members of the polity, a strength exercised in part through less familiar organs of state power, such as presidential EOs and bureaucratic regulation. Combined, these specific aspects of the development and content of affirmative action policy in the United States demonstrate the acuity of taking seriously the political importance of policy instruments as determinants of state–society relations in modern democracies.

Finally, the use of quotas nicely illustrates how policy instruments can have a complex mixture of intentional and unintentional effects. The intended aim of quotas is to de-hierarchize American society and to abrogate the historical injustices associated with the treatment of African Americans and other minority groups in the century between the end of the Civil War and the enactment of the Civil Rights Act in 1964. This goal has been achieved in significant ways—illustrated by the growth of a black middle class, increased percentage of African Americans working in professions, and being educated at elite universities—although inequalities and injustices continue (APSA Task Force on Inequality and American Democracy 2004; Hochschild 1995). Unintentionally, the strategy of employing the quantifiable policy instruments of quotas has forged new divisions resting in part in the very categories—race or ethnicity for instance—which have historically been mobilized to divide Americans internally. It is this latter unintended but fiercely contested consequence of a quantification method of policy instrument which confronts other nations embarking on similar affirmative action initiatives (such as France). The source of this effect lies in the intrinsic character of employing quantifiable policy instruments in public policy.

  
  
  
Acknowledgments
       

Desmond King holds a Leverhulme Trust Major Research Fellowship for his project, "The American State: An Analytical Narrative."

  
  
References
       

 Aberbach, Joel D., and Bert A. Rockman. 2000. In the Web of Politics: Three Decades of the US Federal Executive. Washington, DC: Brookings.
 APSA Task Force on Inequality and American Democracy. 2004. "American Democracy in an Age of Rising Inequality." Perspectives on Politics 2: 651–666. Links  
 Balfour, Lawri. 2003. "Unreconstructed Democracy: W E B DuBois and the Case for Reparations." American Political Science Review 97: 33–44. Links  
 Black, Edwin. 2003. War against the Weak: Eugenics and America's Campaign to Create a Master Race. New York: Four Walls Eight Windows.
 Brinkley, Alan. 1995. The End of Reform. New York: Random House.
 Canon, David T. 1999. Race, Redistricting and Representation. Chicago: University of Chicago Press.
 Dallek, Robert. 1998. Flawed Giant: Lyndon Johnson and His Times 1961–1973. New York: Oxford University Press.
 Davis, Hugh Graham. 2002. Collision Course. New York: Oxford University Press.
 Desrosieres, Alain. 1998. The Politics of Large Numbers: A History of Statistical Reasoning. Cambridge, MA: Harvard University Press.
 Dobbin, Frank, and John R. Sutton. 1998. "The Strength of a Weak State: The Rights Revolution and the Rise of Human Resources Management Divisions." American Journal of Sociology 104: 441–476. Links  
 Dobbin, Frank, John R. Sutton, John W. Meyer, and W. Richard Scott. 1993. "Equal Opportunity Law and the Construction of Internal Labor Markets." American Journal of Sociology 99: 396–371. Links  
 Drake, W. Avon, and Robert D. Holsworth. 1996. Affirmative Action and the Stalled Quest for Black Progress. Urbana: University of Illinois Press.
 Evans, Peter, Dietrich Rueschemeyer, and Theda Skocpol, eds. 1985. Bringing the State Back In. New York: Cambridge University Press.
 Gallagher, Nancy L. 1999. Breeding Better Vermonters: The Eugenics Project in the Green Mountain State. Hanover: University Press of New England.
 Gerstle, Gary. 2001. American Crucible. Princeton, NJ: Princeton University Press.
 Gonzalez, Francisco E. Desmond King. 2004. "The State and Democratization: The US in Comparative Perspective." British Journal of Political Science 34: 193–210. Links  
 Grekul, Jana, Harvey Krahn, and Dave Odynak. 2004. "Sterilizing the 'Feeble-Minded': Eugenics in Alberta, Canada." Journal of Historical Sociology 17: 358–384. Links  
 Hattam, Victoria. 2005. "Ethnicity & the Boundaries of Race: Rereading Directive 15." Daedalus (Winter): 61–69. Links  
 Hochschild, Jennifer. 1995. Facing up to the American Dream. Princeton, NJ: Princeton University Press.
 Hood, Christopher C. 1983. The Tools of Government. London: Macmillan.
 Huntington, Samuel P. 2004. Who Are We? America's Great Debate. London: Simon & Schuster.
 Johnson, Lyndon B. 1965. "Commencement Speech: To Fulfill These Rights." Howard University, Washington, DC, June 4.
 Karl, Barry D. 1983. The Uneasy State. Chicago: University of Chicago Press.
 Kettl, Donald F. 2002. The Transformation of Governance. Baltimore, MD: Johns Hopkins University Press.
 Kevles, Daniel J. 1986. In the Name of Eugenics. Harmondsworth: Penguin.
 King, Desmond. 1995. Separate and Unequal: Black Americans and the US Federal Government. Oxford: Oxford University Press.
 King, Desmond. 2005. The Liberty of Strangers: Making the American Nation. New York: Oxford University Press.
 King, Desmond, and Randall Hansen. 1999. "Experts at Work: State Autonomy, Social Learning and Eugenic Sterilization in 1930s Britain." British Journal of Political Science 29: 77–108. Links  
 King, Desmond, and Rogers M. Smith. 2005. "Racial Orders in American Political Development." American Political Science Review 99: 75–92. Links  
 Krasner, Stephen. 1978. Defending the National Interest. Palo Alto, CA: Stanford University Press.
 Kryder, Daniel. 2000. Divided Arsenal. New York: Cambridge University Press.
 Lascoumes, Pierre, and Patrick Le Galès. 2007. "Introduction: Understanding Public Policy through Its Instruments—From the Nature of Instruments to the Sociology of Public Policy Instrumentation." Governance 20: 1–22. Links  
 Lemann, Nicholas. 1999. The Big Test. New York: Farrar, Straus and Giroux.
 Lieberman, Robert C. 2002. "Ideas, Institutions and Political Order: Explaining Political Change." American Political Science Review 96: 697–712. Links  
 Light, Paul. 1999. The True Size of Government. Washington, DC: Brookings.
 Loury, Glenn C. 2004. "Right and Wrong, Not Black or White." Financial Times Magazine, March 20: 11–12.
 Mayer, Kenneth R. 2001. With the Stroke of a Pen. Princeton, NJ: Princeton University Press.
 Mendelberg, Tali. 2000. The Race Card. Princeton, NJ: Princeton University Press.
 Nettl, J. P. 1968. "The State as a Conceptual Variable." World Politics 20: 559–592. Links  
 Robinson, Greg. 2001. By Order of the President. Cambridge, MA: Harvard University Press.
 Rung, Margaret C. 2002. Servants of the State: Managing Diversity and Democracy in the Federal Workforce 1933–1953. Athens: University of Georgia Press.
 Sabbagh, Daniel. 2003. "Judicial Uses of Subterfuge: Affirmative Action Reconsidered." Political Science Quarterly 118: 411–436. Links  
 Sabbagh, Daniel. 2004. Equal by Law: The Paradox of Affirmative Action. New York: Palgrave Macmillan.
 Skocpol, Theda. 1992. Protecting Soldiers and Mothers: The Political Origins of Social Policy in the United States. Cambridge, MA: Harvard University Press.
 Skowronek, Stephen. 1982. Building a New American State: The Expansion of National Administrative Capacities 1877–1920. New York: Cambridge University Press.
 Skrentny, John D. 1996. The Ironies of Affirmative Action. Chicago: University of Chicago Press.
 Skrentny, John D. ed. 2001. Color Lines: Affirmative Action, Immigration and Civil Rights Options for America. Chicago: University of Chicago Press.
 Skrentny, John D.. 2002. The Minority Rights Revolution. Cambridge, MA: Harvard University Press.
 Smith, Rogers M. 1997. Civic Ideals. New Haven, CT: Yale University Press.
 Sniderman, Paul M., and Thomas L. Piazza. 1993. The Scar of Race. Cambridge, MA: Harvard University Press.
 Stears, Marc. 2002. Progressives, Pluralists, and the Problems of the State. Oxford: Oxford University Press.
 Stuckey, Mary E. 2004. Defining Americans: The Presidency and National Identity. Lawrence: University Press of Kansas.
 Trent, James W., Jr. 1994. Inventing the Feeble Mind: A History of Mental Retardation in the United States. Berkeley: University of California Press.
 Weir, Margaret, and Theda Skocpol. 1985. "State Structures and the Possibilities for Keynesian Responses to the Great Depression in Sweden, Britain and the United States." In Bringing the State Back In, ed. Peter Evans, Dietrich Rueschemeyer, and Theda Skocpl. New York: Cambridge University Press.