Epstein’s Challenge to the Civil Rights Regime*




Every defender of the civil rights regime should read Richard Epstein’s Forbidden Grounds.[1]   Hard as it may be to concede, this book-length treatment of Title VII of the Civil Rights Act of 1964 (and its subsequent legal adaptations) for the first time drives the analysis to the root condition, the antidiscrimination principle itself. In doing so Epstein provides a more than plausible case for a reconsideration, for he proves that the American legal order has not heretofore challenged itself to articulate clearly the full social meaning of the antidiscrimination principle.

“Simple justice” simply conveys this meaning inadequately, for antidiscrimination in principle goes far beyond anything justice demands. In the most obvious case justice can reach no farther than the claims of individuals, however remote or numerous, while the antidiscrimination principle abstracts from persons altogether and deals most comfortably with broad ascriptions – the broader the better – in a way entirely unsuited to judging just claims. This conundrum leads Epstein towards an abandonment of the antidiscrimination principle. One may question, however, whether the principle were ever rightly conceived as a means to adjudicate just claims. Epstein raises the last question but fails to see that a negative response does not entail a negative response to the question of whether the antidiscrimination principle has any role to play in a free political order.

The answer to this question, in turn, leads one to discover in the antidiscrimination principle a new version of the theory of limitedgovernment – i.e., the principle serves to address the historical dilemmas of race and gender discrimination in our society only to the extent that it serves to deny to government these specific means of regulating or adjudicating claims of justice .[2]  Thus, the antidiscrimination principle is properly the limit condition rather than the causa efficiens of a properly established civil rights regime.

Historically, the extension of civil rights protections for individuals from state actions to private actions may seem to belie the claim that the antidiscrimination principle is a form or element of limited government theory. The truly great growth of government power in the realm of civil rights has certainly come in the form of enforcing antidiscrimination standards against private actors. The practical conversion of these standards into “thou shalt not” addressed to persons whether public or private rather than into “government shall make no laws” long since carried the day. This begs the question, however, whether the conversion operated sozein to phainomena (to save the phenomena or appearances). A similar conversion bedevils the very clear example of First Amendment speech protections, however. Justice Black’s First Amendment absolutism yielded ultimately to lemming equivocation, and debates today turn far more usually on questions of whether and how persons, public or private, may speak or pray than on questions of whether and how government may decide such questions. When freedom of speech is viewed as a limit condition, the result is not that anyone may speak or pray unimpeded but only that government may not do the impeding. Conversely, when freedom of speech is viewed as a title to speak or pray, the result is that one may speak or pray only insofar as government confirms the title. The one view limits government’s power to decide; the other empowers government to decide. When antidiscrimination is viewed as a limit condition, the result is not that anyone can be prevented from discriminating but only that government may in no way do the discriminating. Conversely, when antidiscrimination is viewed as a title to be free of discrimination, the result is that one may enjoy the freedom only insofar as government confirms the title. The one view limits government’s power to decide, while the other substitutes the government’s decision as the only substantive meaning of nondiscrimination.

Let us look more closely at the government’s determination of the substantive meaning of nondiscrimination to better evaluate the relation between the current practice of the civil rights regime and, on the one hand, the alternative suggested by Epstein and, on the other hand, the social consequences of the limit condition view.

The most distinct expression of the legal and political determination of nondiscrimination is the theory of disparate impact. According to this theory, discrimination exists within relevant population subgroups if and only if these groups are heterogeneously constructed and the statistical relations of their component elements vary from predetermined reference norms without reference to explicitly formulated exceptions.[3]  From this general theory at least two general consequences follow: (1) no discrimination by race or gender can occur within homogeneous sub-groups, and, (2) no discrimination occurs within heterogeneous groups that are balanced in regard to the reference norms.[4]  (This is a thought experiment which by no means denies the existence of disparate treatment or special case analysis.)

When we apply this general theory to a particular case, we obtain an interesting result, namely, that in an officially nondiscriminating environment it is likely, if not necessary, for excluded members of protected groups to suffer unofficial but very real discrimination. The test case is by no means imaginary, having been supplied in the form of a table proudly submitted to the U.S. Commission on Civil Rights by the Kentucky State Commission on Human Rights in evidence of the success of Kentucky’s then recently-imposed affirmative action plan.



Number and Percent of Black Full-Time Employment

in Kentucky State Government, 1967-1987


Nov. 1967

Nov. 1971

Nov. 1975

Nov. 1977

Nov. 1979

Nov. 1981

Nov. 1983

Nov. 1985

Nov. 1987

Total Full-time Employees










Black Full-time Employees










Absolute Change in Black Employment


+ 132

+ 483

+ 102

+ 582

- 140

-  47

+  85

+  84

Percent Black Employment










Change in Black Share of Employment


+ 0.4%

+ 0.9%

+ 0.2%

+ 0.6%

+ 0.6%

+ 0.1%



Source: Black Employment in Kentucky State Agencies, Kentucky Commission on Human Rights, 1988, page. 5


Inspecting the fourth row in the Table, the reader will perceive at once that population ratio for black people in the State operates as an ironclad quota (4 years running), meaning that neither more nor fewer blacks will be retained in the State government’s work force than called for by the quota! A figure that grew from 5.8 % to 7.2 % in just 6 years remained throughout the succeeding 3 years firmly fixed at 7.3 %. What is the probability against something like that happening randomly, out of a work force of nearly 40,000 people? –something greater by several times than the odds against intelligent life elsewhere in the universe![5]

Incidentally, the fact that the Kentucky State Commission on Human Rights might proudly proclaim these results as evidence of their compliance with Title VII, and that the U.S. Commission on Civil Rights might approvingly receive it, is more than sufficient evidence that quotas enjoy a high level of approval under the Civil Rights Act of 1964. It is likely, however, that that approval depends utterly on the false impression created by the almost universal testimony against quotas.

Thus, the black people who are not already employed in the Kentucky State work force – the excluded members of the protected group – are subjected to approved discrimination. To that degree, it is black folk, not white males, who bear the burden of quotas. This is no surprise. Modest historical sensitivity reveals the necessity. Beginning in slavery and continuing long thereafter, black folk participated in the labor pool at rates far exceeding other subpopulations. Thus, where a rate of 500 or 600 per thousand population would have been high for the average group, a rate approaching 900 per thousand population would have been normal for black folk. In recent years the spread between black folk and others diminished, but it is unlikely that parity has been reached. Accordingly, a quota based on general population ratios, as in Kentucky, actually represents a net loss of jobs for black folk. This job loss is principally in unskilled and blue collar fields, and that helps explain the persistent high unemployment in those areas (and the corollary of welfare subsistence). That was the original protection for labor unions. It also explains the general impression of a displacement of white workers, for that does occur in white collar fields where blacks had been minimally employed. Thus, hiring to a general population level in blue and white collar jobs, while still falling short of historical labor patterns, explains both apparent improvements and high unemployment resulting from discrimination. Reinforce the effect by means of black competition with white women, hispanics, and others, and one has the real picture of the quota regime sponsored by the Civil Rights Act of 1964: nothing more than a regulated version of the open discrimination proposed by Epstein in Forbidden Grounds.

How would a civil rights regime á la Epstein deal with these same facts? According to Epstein, repeal of Title VII would operate as to private parties but not public agencies. In a freedom of contract regime, “Only where the state acts as an employer are there substantive limitations on the kinds of contracts that can be formed.”[6]  It is unclear in his analysis whether theories such as disparate impact may yet be applied in this slimmed-down version of the civil rights regime. If the answer is yes, one could still obtain the Kentucky result.[7]  If the answer is no, one is at a loss to explain how the protections would operate, unless Epstein confines them to tortious actions against government. His discussion in chapter twenty is equivocal, though he does argue that

[i]nsisting on some racial balance in public employment need not be a cynical way to institute rigid quotas .... It helps build legitimacy .... Race, religion, and national origin are useful proxies .... Within vast portions of the public sphere, there will be some carryover of the arguments for rational discrimination along racial or sexual lines.[8]

If government agencies, including many public universities, would look much the same – and perhaps even more – under an Epstein civil rights regime, the real question raised by his analysis must be as to its effect in the private market. Even here, however, there are theoretical difficulties which require close examination. Among these none is more important than the fact that for Epstein a freedom of contract regime is one in which the claims of contract are enforced by the state. His proposal for a repeal of Title VII, accordingly, is actually a proposal to establish secure public protection for discriminatory contracts. It is not, then, the putatively negative proposition (“repeal the Civil Rights Act”) but the daringly positive, proactive proposition, that society should give discrimination the force of law, that deserves closest attention.

Without responding to this wider debate which I have now opened, I can nonetheless demonstrate the implications of Epstein’s position. For I have earlier argued that we should substitute for the world of disparate impact a rigorous application of “state action” theory, including within the concept the action of the court.[9] As Epstein acknowledges the propriety of society refusing to lend its authority to the claims of illicit contracts, such as the contract to murder, I do not need to defend the proposition that the state may legitimately withdraw its protection from certain forms of contract. Accordingly, one needs only to demonstrate that agreements to engage in invidious discrimination qualify as unprotected contracts to pose an alternative, not only to the civil rights but also to Epstein’s version of the freedom of contract regime. Interestingly, the version I have proposed is also a freedom of contract regime, for it maintains that society is better positioned to rely upon freedom of contract once it has made clear what are and what are not acceptable contracts.

As I argued in Runyon, a rigorous application of state action theory not only clears up the problem of discrimination within the private market (by subjecting it squarely to market discipline without an opportunity for public subsidy) but next establishes a forthright prohibition in the public realm, which is far the more important of the two. The argument was straightforward:

[I]t is unjust to empower one private individual to force another private individual into a contractual relation against the latter’s will. That, however, is a misapplication of Runyon and is not its essence, Justice White’s view notwithstanding [it is] essential to bar private or public third-party interference in the right to make contracts. That is the legitimate civil right at which section 1981 was aimed. Congress meant to spell out limits to contracts and to confine their obligations to the contracting parties, as well as to protect entrepreneurial freedom, which alone renders the right of contract meaningful.[10]

The key to this argument was the fact that it does not create a basis for government to coerce unwilling offerors to enter into contracts. Its greatest force, through denial of enforcement, is to undermine contracts which seek to bear on third parties invidiously, insofar as such cartel-like arrangements cannot survive without tacit public subsidy. On the other hand, private, two-party agreements (logically speaking) may well be discriminatory and will remain effective for as long as the parties maintain their motivation to sustain the agreement.[11]  That such a result is possible is attested by the prevailing racial characteristics of marriage. That a pervasive tendency in this regard may nonetheless be altered by private dealing, secured against meaningful third-party interference, is attested by the occurrence of intermarriage.

In the Kentucky test, the limited government freedom of contract regime based on denying state sanction to all invidious discrimination would result in elimination of the official discrimination which now results. For any individual could raise a meaningful objection to public discrimination without it being sufficient to point to mere numbers to rebut the allegation. Moreover, in private markets the intrusive hand of government would be withdrawn precisely as Epstein wishes but without creating a state-sanctioned private discrimination.[12] Rather, the general refusal to enforce discriminatory contracts would operate in such a way as to bear most significantly in the case where the social significance of the arrangement has greater importance – where the scale is greater. “The private school system would thus be opened because the cost of operating on unenforceable contracts would be too high, while the baby-sitting contract would never be called into question.”[13]

Epstein’s retort is not compelling, but it should be stated:

In some cases there may be a question of drawing the line as to whether certain decisions have been made by the state or by private parties: a state collective bargaining statute that confers on a union a legal monopoly is one such example. But as a first approximation most businesses operate not under the protection of a legal monopoly but only with the ordinary protections that are afforded by the law of property, contract, and tort. To say that the receipt of these legal protections converts private decision into state action does away with any distinction between state and private action, and thus undermines the same legal regime of private rights that the Civil War Amendments sought to protect.[14]

Beyond the fact that this argument carries a normative rather thanan analytical force, it strikes me that it misses the point. For Epstein has maintained that most deals in the private market are stable, not requiring to be supported by the appeal to state enforcement or, rather, not generally being called into question before state officers. Hence, not every private act becomes state action even on the theory he opposes! Only unstable private acts which can be resolved and upheld only with recourse to state power and authority become state acts. In that light, nothing can be more reasonable than for the society to consider in advance which kinds of instability it wishes the state to invest its authority in settling and which it does not.[15]  That is the very essence of limited government.

I believe this contrast in freedom of contract regimes to be far more powerful than the contrast between Epstein’s freedom of contract regime and the civil rights regime. The prevailing climate of opinion, however, makes Epstein’s move the more compelling, for he accomplishes with it the needed reappraisal of the goals of the civil rights regime. The book constitutes a challenge to defenders of the civil rights regime to take up the one question they have consistently refused to entertain namely, what is the relation between a freedom of contract regime and a civil rights regime. In fact, the legitimacy of civil rights legislation will remain doubtful unless and until an adequate response can be given to that question. In that sense, Epstein has not undermined the civil rights regime so much as he has demanded that it complete the construction of that edifice on the only foundation that can enracinate it within the American political order.

I have emphasized our differences on this subject, but these differences should not obscure the forthright agreement we share on what is, perhaps, the most important sentence in Forbidden Grounds, namely, that “[t]he refusal to deal for any reason lies at the root of a system of a freedom of contract, itself the centerpiece of any common law order based on the autonomy principle.”[16] It is here that defenders of the civil rights regime must engage Epstein’s argument; failing to do so will be to fail utterly either to defend or to transcend the civil rights regime.



* Published in the San Diego Law Review, vol. 31, no. 1 (Winter 1994): 57-66.


** Dean, James Madison College, Michigan State University.




[2] In fairness to Epstein, one must acknowledge that he approached this view of the problem in the essay which anticipates Forbidden Grounds, Two Conceptions of Civil Rights, in REASSESSING CIVIL RIGHTS 38-59 (Ellen Paul et al., eds., 1991). There, Epstein argued that a “libertarian conception” of civil rights could be contrasted with the more recent, antidiscrimination version of civil rights. His libertarian or “individual” conception of civil rights approaches but is not quite the same as the “limited government” concept employed here. Epstein did not repeat this contrast in the same forceful terms in his book-length discussion, perhaps because he could discern difficulties in the oversimplification caused by the contrast as he had made it originally.


[3] This general and abstract statement, I submit, covers every conceivable qualification relative to work forces and “qualified” work forces any court either has or can dream up.


[4]  This means, obviously, that within a group defined by race, discrimination by gender can occur and vice versa, but discrimination by gender within a group defined by gender cannot occur, and so on.





[6] EPSTEIN, supra note 1, at 4.  It is at least logically incorrect, parenthetically, to say that an unprotected contract is therefore a “forbidden contract.”


[7] He reasonably excludes voluntary affirmative action by government, leaving quotas as the likeliest substantive limitation if tort remedies are not invoked. EPSTEIN, supra note 1, at 9.


[8] EPSTEIN, supra note 1, at 424-25.


[9] William B. Allen, Keynote Address: Let’s Re-Do Runyon: Questions to Guide Justice White, 41 RUTGERS L. REV. 893 (1989) [hereinafter referred to in the text as Runyon].


[10] Id. at 896.


[11] Here again, Epstein steps from certain principle to uncertain application: “With freedom of contract, therefore, the legal system judges only those contractual offers that are made and accepted. Thereafter, competition within the economic system, and not government fiat, determines which hiring strategy is superior.” EPSTEIN, supra note 1, at 290. In light of his reasonable attack on government-enforced segregation, Epstein ought to be able to see that offering judicial protection for such practices – while not overriding the market – steps onto that slippery slope which descends into “political market” attempts to defend “legitimate” contracts against market discipline, which is not only the real story of Jim Crow but perhaps of every significant form of regulatory intervention into the market.


[12] The defect of the restrictive covenant, which Epstein would allow, is that it binds a contractee to a performance which can not be disciplined. Thus, in forming the contract, the incentive to concede to a seller this evanescent utility is especially high, since subsequently the seller cannot retain the advantage gained saved by undergoing the opportunity costs and real costs of legal enforcement.


[13] Allen, supra note 9, at 904.


[14] EPSTEIN, supra note 1, at 131.


[15] Whereas Epstein thinks it should make no difference to the state:

There is, moreover, no reason to take into account the so-called negative externalities of the practice [private preference/quota contract schemes], for with ordinary two-party contracts the only externalities of any legal relevance are the threat of force against strangers and the use of monopolistic practices, both of which are as remote from affirmative action programs or quotas as from any other form of employer choice.  EPSTEIN, supra note 1, at 415.


[16] EPSTEIN, supra note 1, at xii.

Back to top

Back to List of Articles