What Indirect Affirmative Action Can Do

III.A. Indirect Affirmative Action Is Not Necessarily Unjust

I will start by arguing that indirect affirmative action is not necessarily unjust. This is a matter of defending indirect affirmative action against objections. Obviously, I cannot, in this paper, consider every conceivable objection to indirect affirmative action. Instead, I will consider the most prominent objections put forward against affirmative action, partly because these have played the largest role in discussions of affirmative action, partly because the frequency with which they appear suggests that they may be considered to be the strongest objections to affirmative action. These are the stigma objection, the balkanization objection, the merit objection, the publicity objection, and the mismatch objection. I will argue that these fail to establish that indirect affirmative action is necessarily unjust. I should clarify immediately that I do not aim to show that indirect affirmative action can never be unjust. I do not think that is true. What I instead aim to show is that these objections fail to show that it is necessarily unjust to pursue indirect affirmative action.

The stigma objection. A common objection to affirmative action is that it stigmatizes its recipients. Here is Cohen, dramatically putting forward this objection:

If some demon had sought to concoct a scheme aimed at undermining the credentials of minority scholars, professionals, and students, to stigmatize them permanently and humiliate them publicly, no more ingenious plan could have been devised than the system of preferences now defended as a social need and great favor to minorities.²⁹

The quote clearly illustrates the gist of the stigma objection: affirmative action leads to stigmatization of recipients (and other members of the recipient group). Indeed, it may even threaten the self-esteem of the recipients, e.g., because they start to question whether they got hired because of their race or gender.³⁰

To discuss this objection, it may be useful to focus on a particular example of indirect affirmative action. So, suppose we provide a benefit in university admissions to applicants whose parents did not go to university, knowing that members of a particular disadvantaged group are significantly more likely to have parents who never went to university.³¹ It is, in the end, obviously an empirical question whether this indirect affirmative action policy would lead to stigmatization of its recipients. But I think there is some reason to think that, at least sometimes, it would not. To stigmatize an affirmative action recipient, the stigmatizer would need to know where to direct their stigma. With facially non-neutral policies, especially when these single out groups with visible characteristics, such as race or gender, it may be fairly easy to identify potential affirmative action recipients and stigmatize accordingly. But it is not visible in the same way whether a person’s parents went to university or not. And stigmatizers may not be particularly well-informed: they might not know that there is a high correlation between membership of a protected group and having parents who never went to university. If so, identifying and stigmatizing potential recipients will be much more difficult.³² Indeed, they might not even be aware that the policy amounts to affirmative action. This will most likely be the case when the policy can be rationalized independently of the benefits it provides to the protected group.³³ And this seems to be the case in this example, the rationale being that it will be much more difficult to be admitted to university if your parents never went to university than if they did. So, although it is ultimately an empirical question, there is reason to think that indirect affirmative action, precisely because it is facially neutral, will not necessarily lead to stigmatization of its recipients.

Even if we assume that an indirect affirmative action policy would lead to stigmatization of its recipients, it is not clear why this should necessarily make the policy unjust. For one thing, there may be more stigmatization, including feelings of being stigmatized, in the status quo (without affirmative action). If so, pointing to the policy’s stigmatizing effect does not seem to be a good argument for the policy being unjust. For another thing, it may be, instead, that it is not the indirect affirmative action policy which is unjust, but the stigmatizing response to the affirmative action policy. For instance, it may be that by stigmatizing a potential recipient, the stigmatizer wrongs the potential recipient by disrespecting their individual agency³⁴ or by committing an epistemic injustice against them.³⁵ If the stigmatizing response is unjust, then pointing to that stigmatizing response is not a good objection to the policy, especially not if potential recipients would prefer the presence of policy and stigma to the absence of both.³⁶

The balkanization objection. In Shaw v. Reno,³⁷ Justice O’Connor put forward the balkanization objection:

Racial classifications of any sort pose the risk of lasting harm to our society. They reinforce the belief, held by too many for too much of our history, that individuals should be judged by the color of their skin. Racial classifications with respect to voting carry particular dangers. Racial gerrymandering, even for remedial purposes, may balkanize us into competing racial factions.³⁸

Rather than leading to an integrated society, the balkanization objection asserts that affirmative action leads to antagonism between groups; indeed, that “interventions promoting racial integration can become a locus of racial conflict.”³⁹ This objection “attends to the concerns of the dispreferred” and asserts that affirmative action results in a lack of social cohesion in society.⁴⁰

There are similarities between the balkanization objection and the stigma objection in the sense that what is objectionable comes by as a result of how people respond to the affirmative action policy. In the stigma case, because majority individuals stigmatize potential affirmative action recipients. In the balkanization case, because majority individuals respond with animus or something similar to the members of the recipient group, leading to conflict between groups. This similarity is useful in the present context as it suggests that a response to one may also be a response to the other. And we saw in the discussion of the stigma objection that there are instances of indirect affirmative action that are not likely to cause stigmatization. This was the case with the example of providing a benefit in university admissions to applicants whose parents did not go to university. If that was true in the stigmatization case, it is likely true in the balkanization case as well, precisely because both are a result of how advantaged groups respond to the policy. This—that indirect affirmative action may not lead to balkanization—seems to be in line with discussions in the US Supreme Court. As Siegel points out,

Antibalkanization understands that race-conscious, facially neutral interventions may promote social cohesion by promoting equal opportunity, as Justice Kennedy demonstrates in Parents Involved and Ricci when he discusses permissible forms of race-conscious, facially neutral action by administrators siting school districts and employers complying with the disparate impact provisions of federal employment discrimination law.⁴¹

Of course, this is not to deny that some indirect affirmative action policies could lead to balkanization.⁴² It is just to say that there is good reason to believe that some will not. And that suffices for our purposes.

Even setting that aside, there are clear limits to the balkanization objection. First, there might be balkanization in the status quo (without affirmative action), such that there would be less balkanization overall with indirect affirmative action than without. If so, pointing to the policy’s balkanizing effect does not seem to be a good argument for the policy being unjust (as I similarly pointed out in response to the stigma objection). Second, advantaged groups could always threaten that an intervention to mitigate injustice would lead to social discordance. After all, the discordance would follow from their response to the intervention. But this clearly cannot always be a good objection to an intervention. If rich people complained that a taxation scheme would lead to balkanization between rich and poor people, we would not think that would be a good objection to having a taxation scheme. And the main reason for this is that they would make it true that there would be balkanization: it would be a result of how they responded to poor people as a result of the taxation scheme, and they could have responded differently. As Cohen has taught us, such a response by rich people would not be a good argument for the scheme being unjust.⁴³ If so, we should be able to say the same in response to the balkanization objection, at least in relation to some indirect affirmative action schemes (given the similarities between the tax case and the no-parent university case, the latter indirect affirmative action scheme seems to be such a case). Again, it may be the response to the policy that is unjust, and not the policy itself.

The merit objection. Another prominent objection to affirmative action, the merit objection, says that affirmative action violates the entitlement of the best qualified. If we pursue affirmative action in, say, hiring or in admissions to university, and hire or admit a minority candidate who is less qualified than a majority candidate, we wrong the majority candidate by failing to give them that to which they are entitled in virtue of being the best qualified, to wit, the job or the spot. This is, in essence, the merit objection.⁴⁴

It is important first to make certain that we do not confuse the view that the best qualified should be hired with other views with which it is correlated. If we fail to do so, we may give too much credit to the merit objection. We must distinguish the view that the best qualified should be hired from the view that one should not disappoint applicants’ reasonable expectations.⁴⁵ If hiring usually takes place in accordance with merit, people may reasonably expect, when they apply for a post, that the post will be filled according to merit. So if we do not fill the position according to merit, we disappoint the applicants’ reasonable expectations. But this does not speak in favor of the view that the best qualified must be hired, as opposed to the view that we should not disappoint reasonable expectations. If indirect affirmative action were announced in advance, it would not disappoint reasonable expectations, and it would thus not be unjust for this reason. We must also be careful not to suppose that the best qualified must be hired because we are afraid that insufficiently qualified candidates will be hired.⁴⁶ We may simply restrict indirect affirmative action to those candidates who are sufficiently qualified.⁴⁷ The two factors we have just considered—reasonable expectations and (in)sufficient qualifications—may provide an error theory for why the view that the best qualified should be hired seems to have much force. But it is important that we distinguish these considerations; they do not provide support to the merit objection.

I suppose this takes most of the force out of the objection. But even if we assume this is not the case, it is not even clear that indirect affirmative action entails that the best qualified isn’t hired. Two distinctions are relevant here. First, merit can be understood in a narrow sense or in a broad sense. Merit in the narrow sense is technical skills, e.g., how many clams you can collect in an hour. Merit in this sense is unaffected by whether you are a member of a minority group. On a broad understanding of merit, such group membership may be a qualification. As Dworkin says, “there is no combination of abilities and skills and traits that constitutes ‘merit’ in the abstract … If a black skin will, as a matter of regrettable fact, enable another doctor to do a different medical job better, then that black skin is by the same token ‘merit’ as well.”⁴⁸ If a neighborhood with Black people is medically underserved, being a Black doctor may relevantly count as merit.⁴⁹ If so, an indirect affirmative action scheme may lead to hiring the best qualified candidate in the broad sense. And there is no reason to believe that the narrow understanding of merit is the most appropriate one in all situations.

Second, there is a difference between being most qualified, individually speaking, and being most qualified, collectively speaking.⁵⁰ Suppose you must work in a team. Whether you are the most qualified, individually speaking, is independent of who else will be working in the team. Whether you are the most qualified, collectively speaking, is dependent on who else will be working in the team. If the others on the team are men, the team might benefit from the diversity of you being a woman, such that you are the most qualified in the collective sense (but not necessarily in the individual sense). Indirect affirmative action may lead to hiring the most qualified, collectively speaking. And there is no reason to believe that the individual understanding of merit is the most appropriate in all situations. If so, the merit objection fails to show that indirect affirmative action is necessarily unjust.

The publicity objection. A common charge against affirmative action is its lack of publicity.⁵¹ The charge is that it violates what Lippert-Rasmussen refers to as the liberal publicity constraint: “the principle that public policies, rules, etc. must be stated openly in such a way that they are accessible to the public.”⁵² This makes affirmative action unjust. Two implicit premises are assumed here: that anything that violates the publicity constraint is unjust, and that affirmative action violates the publicity constraint. I address these premises in turn.

Clearly, there must be exceptions to the liberal publicity constraint. Few would say that the work of intelligence agencies is unjust because it is not accessible to the public.⁵³ What explains this exception? Presumably that it would undermine the work of intelligence agencies if their work was accessible to the public. And since the work is important, that would be undesirable. But one might then point out that the “work” of some indirect affirmative action policies is important as well, especially when there is much injustice in society. And it might also be that the work of some of these policies would be undermined if it was accessible to the public. So why couldn’t indirect affirmative action also sometimes be an exception to the liberal publicity constraint?

And even if indirect affirmative action must satisfy the liberal publicity constraint to not be unjust, there is nothing in indirect affirmative action as such that precludes stating openly why the policy is pursued, e.g., to secure equality of opportunity between members of disadvantaged and advantaged groups. While it may be true that some affirmative action policies that have been pursued have failed to satisfy the liberal publicity constraint, that cannot be an objection to indirect affirmative action as such. Those who pursue the policies could state the aim to the public, in which case they would not be unjust for publicity reasons. Perhaps the worry is that affirmative action will not work as well if stated publicly as when done in secret.^54,55 But this does not show that indirect affirmative action is unjust. It cannot be an objection to a just policy that an unjust policy would be more effective. And if indirect affirmative action must satisfy the liberal publicity constraint to be just, the objection would have precisely this form: it would object to the public indirect affirmative action policy (the just one) that the secret indirect affirmative action policy (the unjust one) would be more effective. But that is neither here nor there if we do not want to pursue unjust policies. Thus, the publicity objection fails to show that indirect affirmative action is necessarily unjust.

The mismatch objection. Let us consider a final objection. According to this objection, affirmative action is unjust because it is underinclusive and overinclusive in relation to beneficiaries. When it comes to beneficiaries, affirmative action is underinclusive because it does not benefit some who should be benefited (disadvantaged minority individuals who will not be positioned to take advantage of the opportunities provided by affirmative action, and majority individuals who will have faced relevantly similar injustices as minority individuals), and it is overinclusive because it benefits some who should not be benefited (advantaged minority individuals who will be positioned to take advantage of the opportunities).⁵⁶

While it may be a good objection to some affirmative action schemes, the mismatch objection cannot establish that indirect affirmative action is necessarily unjust. First, a mismatch must be relative to a particular aim. There is no fact of the matter about the mismatch involved in a policy unless we know what the policy’s aim is. And we can imagine many different aims that could be pursued in indirect affirmative action policies. If our aim is to mitigate patterned inequality, then overinclusiveness may not be a worry, as Eidelson illustrates in the context of anti-discrimination norms:

If we understand anti-discrimination norms as interventions to reduce patterned inequality, then particular individuals are granted rights under those norms for the simple reason that their fates are bound up with the larger social problem under attack. In other words, it happens that getting them into more favorable positions would be an improvement in the pattern. There are undoubtedly some other individuals who are just as badly off and yet receive no similar aid, and there are surely others who are morally wronged by the decisions made about them, in the same domains of decision-making, and yet are given no legal recourse. Excluding those others from the protection of anti-discrimination laws is justified, from this point of view, not because they are less deserving of help or have not suffered as serious a wrong, but simply because helping them would not have the social benefit of undermining patterned inequality.⁵⁷

Second, most, if not all, policies in societies as we know them will involve mismatches. The penal system also involves a mismatch—some guilty people go free, and some non-guilty people do not—but that would not necessarily lead us to the conclusion that the penal system is unjustified.⁵⁸ This shows that the relevant concern is not whether a policy involves a mismatch, but whether it involves less or more of a mismatch than relevant alternatives.⁵⁹ If there is a relevant alternative to indirect affirmative action which involves less of a mismatch, then we may conclude that the affirmative action policy is unjust, all else equal. But it is highly unlikely that indirect affirmative action will always be worse than a relevant alternative policy in terms of mismatch, especially once we remember that “when faced with a mismatch involving an affirmative action program, it is always possible that it could be revised in such a way that the relevant mismatch is reduced.”^60,61 And, finally, sometimes the mismatch in a given indirect affirmative action policy may simply be a result of society being extremely unjust—the mismatch is a result of where the injustice has “placed” people in society—in which case what we should object to may be the prior injustice, and not the indirect affirmative action policy.

III.B. Reasons To Pursue Indirect Affirmative Action

We have seen that five prominent objections to affirmative action fail to establish that indirect affirmative action is necessarily unjust. Obviously, there could be other objections, as I said, but my responses to the objections discussed above may be useful in answering these objections as well. In any case, awaiting and setting aside such possible objections, we can (tentatively) conclude that indirect affirmative action is not necessarily unjust. This does not yet give us a positive reason to pursue indirect affirmative action—it simply establishes the negative point that it would not necessarily be unjust. So, in this section, I will point to some positive reasons to pursue indirect affirmative action. I will not spend too much time on this, partly because I will rely on arguments put forward in favor of (standard, direct) affirmative action, partly because I am more interested in exploring the question I will take up in the next section, to wit, the relationship between direct and indirect affirmative action. But first it is important to establish that there is sometimes good reason to pursue indirect affirmative action.

What I will do is to show that what I consider to be two of the strongest arguments in favor of affirmative action—the equality of opportunity argument and the integration argument—also give us reason to pursue indirect affirmative action. I do not have the space to give a full defense of these arguments, so I will rely on the assumption that it is important to secure equality of opportunity and integration and that these values give us reason to pursue direct affirmative action.⁶²

First the equality of opportunity argument. This argument says that we have reason to pursue affirmative action because this brings us closer to a level playing field—a playing field in which minority and majority individuals have equal opportunities (or at least less unequal opportunities). Sher sums up this argument,

the key to an adequate justification of reverse discrimination [affirmative action] [is] to see that practice, not as the redressing of past privations, but rather as a way of neutralizing the present competitive disadvantage caused by those past privations and thus as a way of restoring equal access to those goods which society distributes competitively.⁶³

Now, we must distinguish between formal and substantive equality of opportunity. In Rawls’s words, the latter requires that “those with similar abilities and skills should have similar life chances,” and the former requires that “careers [be] open to talents.”⁶⁴ Proponents of the equality of opportunity argument usually have substantive equality of opportunity in mind; I follow suit.

Just as direct affirmative action may lead to less inequality of opportunity, indirect affirmative action may do so as well. Consider, again, the indirect affirmative action policy of providing a benefit in university admissions to applicants whose parents did not go to university, knowing that members of a particular disadvantaged group are significantly more likely to have parents who never went to university. Since we know that (lack of) cultural capital travels from parents to children, those whose parents did not attend university can be expected to have worse opportunities to being admitted to university than those whose parents did attend university.⁶⁵ And when we add that members of a disadvantaged group are significantly more likely to have parents who did not attend university, pursuing this indirect affirmative action policy can be expected to improve the opportunities of members of this disadvantaged group vis-á-vis members of advantaged groups. I do not mean to suggest that indirect affirmative action will always promote substantive equality of opportunity. But because of how disadvantage tends to cluster,⁶⁶ there is good reason to believe that indirect affirmative action will sometimes (indeed often) promote substantive equality of opportunity. This suffices for my purposes.

The second argument I will consider is the integration argument—an argument put forward by Anderson. As she points out, “Americans live in a profoundly racially segregated society. De facto racial segregation unjustly impedes socioeconomic opportunities for disadvantaged racial groups, causes racial stigmatization and discrimination, and is inconsistent with a fully democratic society.”⁶⁷ Affirmative action is justified as a way of securing integration—non-segregation and non-stigmatization—in society.

Indirect affirmative action may, like direct affirmative action, promote integration or mitigate the disadvantages of segregation. To illustrate, let’s focus on the case of providing a benefit to applicants to medical school whose parents did not go to university, knowing that Black people are significantly more likely to have parents who never went to university. This indirect affirmative action policy may promote integration and mitigate the disadvantages of segregation. As Anderson says, “For professional schools, affirmative action helps remedy the severe deficit residents of segregated neighborhoods suffer in access to professional services. Black physicians are far more likely than white physicians to locate in underserved minority neighborhoods and serve far more black, Latino, and Medicaid patients, even after controlling for their location.”⁶⁸ If Anderson is right, indirect affirmative action could promote non-segregation—in better integrating Black people in the medical profession and the higher echelons of society more generally—and mitigate the disadvantages of segregation—the lack of adequate access to medical services for Black people. Of course, much more could be said here. But if Anderson is correct that direct affirmative action promotes integration, we should also expect this to be true of indirect affirmative action. So there is, it seems, reasons from equality of opportunity and integration to sometimes pursue indirect affirmative action.⁶⁹

III.C. The Strategic Usefulness Of Indirect Affirmative Action

We have seen that there is sometimes good reason to pursue indirect affirmative action. But these reasons likely also give us reason to pursue direct affirmative action. What should we think of the relationship between direct and indirect affirmative action? In this section, I argue that, in an inegalitarian political climate like the current one in the US, indirect affirmative action is, for the egalitarian, particularly strategically useful vis-á-vis direct affirmative action when it comes to achieving egalitarian effects,⁷⁰ and that, in such circumstances, it is at least sometimes permissible for the egalitarian to pursue indirect affirmative action as a form of egalitarian gamesmanship.

To see what I am after, it will be useful to start by turning our attention to a discussion by Eidelson of what he calls second-order discrimination. Second-order discrimination is, as it were, the opposite of paradigmatic indirect affirmative action: “a person commits second-order discrimination on the basis of P when she discriminates on the basis of P in adopting a rule or decision to discriminate on the basis of Q in some other dimension of treatment.”⁷¹ For instance, the racist second-order discriminator does not discriminate directly on the basis of race, because that is prohibited (as a form of direct discrimination); instead, they discriminate on the basis of some other trait which is correlated with race to get much of the effect they would get from discriminating directly on the basis of race. The following example of second-order discrimination is instructive:

In the late nineteenth century, Anatole French satirized French pride in the “majestic equality of the laws, which forbid rich and poor alike to sleep under the bridges, to beg in the streets, and to steal their bread … It is true that the law against sleeping under a bridge does not discriminate on the basis of wealth, since it applies to “rich and poor alike.” But the glaringly disproportionate burden that this law places on the poor is powerful evidence regarding how the law came about. It suggests discrimination on the basis of wealth in the dimension of how different people’s interests are valued in making the laws … France’s commentary alleges discrimination on the basis of wealth in the law’s genesis … when the state applies the formally wealth-neutral law, it does not discriminate against the poor on the basis of wealth; it discriminates only on the basis of where one sleeps. Nonetheless, the state may be guilty of second-order discrimination on the basis of wealth.⁷²

In this case, let us suppose,⁷³ the French government did not want homeless people to sleep under bridges. But it was not allowed to discriminate directly based on wealth. So it adopted a rule—forbidding everyone, irrespective of wealth, from sleeping under bridges—which would primarily negatively affect homeless people. Well-off people would not need to sleep under bridges to begin with. It adopted a facially neutral rule to disproportionately disadvantage homeless people. As Eidelson points out, second-order discrimination “often involve[s] some form of subterfuge or gamesmanship on the part of the discriminator.”⁷⁴ The second-order discriminator games the system to achieve the effect—in the above case, that homeless people do not sleep under bridges—that they cannot achieve by discriminating directly on the basis of wealth (because such actions are prohibited).

What makes such second-order discrimination useful is its facial neutrality when it comes to protected characteristics. In a politically egalitarian climate, it can be (more) effective for the racist to second-order discriminate rather than “first-order” discriminate. It would be unjust to do so, of course—racism, also when hidden, is unjust—but the racist effects that the racist is after will often be better realized second-orderly than first-orderly in such an egalitarian political climate.

And now we return to the current political climate in the US which is not egalitarian but highly inegalitarian. Not only is direct affirmative action prohibited, as per SFFA; even if we set that aside, in this political climate—which includes President Trump’s executive orders regarding DEI policies—it will be strategically unwise for the egalitarian to adopt direct affirmative action policies. As an egalitarian, you do not, for fear of repercussions, want to be branded as someone engaged in affirmative action. It is in such political circumstances, I think, that indirect affirmative action has a strategic advantage vis-á-vis direct affirmative action. And the reason for this is precisely that indirect affirmative action is facially neutral; it does not single out a protected trait. Indeed, in such inegalitarian political circumstances, indirect affirmative action can be used as a form of egalitarian gamesmanship. By this, I mean that the egalitarian—wanting to secure justice for minorities—can game the inegalitarian political system by adopting rules which are facially neutral but which disproportionately advantage minority individuals. In this way, the egalitarian can secure some of the egalitarian effects justice requires without being branded as someone engaged in (direct) affirmative action. For instance, by intentionally selecting an admissions rule which is facially neutral but disproportionately advantages members of a racialized minority group. In this way, indirect affirmative action may be strategically useful for the egalitarian in a way that direct affirmative action is not in a political climate like the current one in the US, and this precisely because of its facial neutrality.

I recognize that this talk of egalitarian gamesmanship may sound controversial to some. So let me clarify what I am and will be arguing. Two points. First, I’m exploring the relationship between direct and indirect affirmative action. Particularly, I’m arguing that, in an inegalitarian political climate like the current one in the US, the facial neutrality of indirect affirmative action is precisely what makes it strategically useful vis-á-vis direct affirmative action when it comes to achieving egalitarian effects. Second, I want to make a preliminary case for the permissibility of pursuing indirect affirmative action as a form of egalitarian gamesmanship in such a politically inegalitarian climate. And while the two are closely related, one may agree with the first point—that, for the egalitarian, indirect affirmative action has a strategic advantage vis-á-vis direct affirmative action in a politically inegalitarian climate because of its facial neutrality—while disagreeing with the second point, e.g., because one might think that the secrecy or deceit involved in using indirect affirmative action as a form of egalitarian gamesmanship makes it impermissible. Even if I only get you on board with the first point, that is still significant: we now know more about the relationship between direct and indirect affirmative action—something we did not know much about to begin with. Having focused primarily on the first point above, then, let me now focus on the second point. This should also help to make clearer how indirect affirmative action can be used strategically by the egalitarian, thereby further substantiating the first point.

As I do not have the space to make a full defense—what I say is meant as a springboard for further discussion of the usefulness of indirect affirmative action—I think it will be useful to make my preliminary case by responding to three negative reactions that I expect some might have: (1) pursuing indirect affirmative action as a form of egalitarian gamesmanship is deceptive and therefore impermissible; (2) even if we assume it doesn’t violate the law, it at least violates the spirit of the law and is therefore impermissible; (3) if second-order racist discrimination is impermissible, why isn’t indirect affirmative action, egalitarian gamesmanship-style, impermissible as well?

Let me start with (1). Pursuing indirect affirmative action as a form of egalitarian gamesmanship requires some lack of transparency: to effectively game the system in a politically inegalitarian climate like the current one in the US, the egalitarian wants to hide the real intention behind their policy.⁷⁵ For instance, they want the admission rule to look like need in a protected-characteristic-neutral way (e.g., a benefit in admission to those whose parents never went to university), although the rule is adopted to disproportionately advantage minority applicants. This lack of transparency about why one is adopting a given rule might remind some of Williams’s criticism of government house utilitarianism, where policy makers use utilitarianism as their guide for making policy, while keeping this secret to the public to maximize total happiness (given the assumption that total happiness might be reduced if citizens knew that their policy makers were utilitarian).⁷⁶ They hide the real reasons for their policy choice, as it were. And this takes us back to the liberal publicity constraint discussed earlier, according to which “public policies, rules, etc. must be stated openly in such a way that they are accessible to the public.”⁷⁷ The egalitarian pursuing indirect affirmative action as a form of egalitarian gamesmanship hides their real reason for adopting a given rule, thereby deceiving people and violating the publicity constraint. This makes it impermissible, the objection goes.

A few responses on behalf of the egalitarian. First, we should ask: why is publicity important? A common answer is that justice is a cooperative enterprise, and that cooperators must know that others are cooperating—to know that they are not being unfairly exploited—and this requires publicity.⁷⁸ But in our case—the inegalitarian political climate—the egalitarian already knows that the inegalitarian others are not cooperating—indeed, that is why they have to pursue indirect affirmative action as a form of egalitarian gamesmanship. The inegalitarian others do not abide by justice. So this does not seem to provide a reason for why the egalitarian should satisfy publicity. Another common answer to “why publicity?” is autonomy, where this involves “a willing identification with the social constraints to which one is subject.”⁷⁹ But minority individuals may not have much autonomy in this sense to begin with in the inegalitarian political situation. Indeed, protecting their autonomy, including in the long term, would seem to sometimes speak in favor of non-publicity. And we can use a similar response to a third common answer to “why publicity?”, that publicity is valuable because it secures a valuable form of community between citizens, which involves a common pursuit of shared ends:⁸⁰ presumably, there isn’t such a community to begin with in the circumstances we consider. These responses suggest a more general reply. While liberal publicity might be a reasonable constraint in ideal justice—conditions with full compliance (or at least where individuals have a sense of justice and a willingness to abide by just rules as long as they are assured that others are likewise inclined and act accordingly)—it does not follow that it is a reasonable constraint when some are unwilling to comply with the demands of justice.⁸¹

Second, a comment specifically on government house utilitarianism. This criticism is, indeed, common, but I wonder how much of its intuitive appeal is due to it being government house utilitarianism which is debated. Consider an instance of government house egalitarianism, where citizens, including racist citizens which compose a large majority, are being told and believe that government is based on inegalitarian, racist principles when, in fact, the policies are based on egalitarian principles. This strikes me as much less intuitively objectionable—indeed, this seems to me to be a reasonable way of achieving justice under non-ideal conditions. Third, even if we set these points aside, publicity is, in any case, just one consideration. As Enoch says, “perhaps transparency matters, even intrinsically. It is just that even if it does, in politics it is likely to be outweighed by suffering almost every single time.”⁸² We can do with something less strong in our case (it doesn’t have to be true that suffering almost every single time outweighs publicity). Even if publicity is one consideration for the egalitarian, mitigating the injustice from which minority individuals suffer is another consideration—indeed, a lot of suffering might be at stake for minority individuals—and it seems unbelievably strong to say that the former always outweighs the latter.⁸³ If so, publicity considerations will likely not always make it impermissible for the egalitarian to pursue indirect affirmative action, egalitarian gamesmanship-style.⁸⁴

Let’s turn to the second worry: even if pursuing indirect affirmative action in this way does not violate the law, it at least seems to violate the spirit of the law, and this makes it impermissible for the egalitarian to do so. In short: the prohibition on direct affirmative action also extends, at least in spirit, to indirect affirmative action. A few reactions to this concern. First, remember that there is a relevant difference between direct and indirect affirmative action: the former is not facially neutral—it picks out minority groups for special treatment—but the latter is. If (part of) what makes direct affirmative action prohibition-worthy is its facial non-neutrality, then it does not seem to extend (straightforwardly) to indirect affirmative action.⁸⁵ Of course, it might be that there is a reason for the prohibition that could extend, somewhat naturally, to indirect affirmative action, but the point is that this reasoning would then have to be further developed.

Second, in any case and since I’m not a legal scholar, let’s simply suppose that the reasoning could be developed such that it would hit the egalitarian pursuing indirect affirmative action, egalitarian gamesmanship-style: the egalitarian would act against (the spirit of) the law. Would that make it impermissible for the egalitarian to do so? Not necessarily, I think. Of course, this question deserves a paper-length treatment, or much more, on its own—as the debate on (un)civil disobedience shows⁸⁶—but here are a few remarks to mitigate this illegality worry. This is basically a question of whether it is always morally required to obey (the spirit of) the laws. In this context, Wellman usefully reminds us, it is important to distinguish between political legitimacy and obligation. Political legitimacy concerns the state’s moral right to coerce or create legally binding rules, and citizens’ correlative lack of a right not to be coerced. Political obligation concerns citizens’ moral obligation to obey the laws. Whereas the latter is about what a citizen is obligated to do, the former is about what the state is permitted to do.⁸⁷ Following this, it might be that a state is legitimate—in the sense that it may coerce citizens—but that citizens, or some of them, do not have a duty to obey the laws, for instance, because there is widespread injustice. The situation the egalitarian we consider is in might be such a situation: even if the injustice against minority individuals does not make the state illegitimate, it is not obvious that the egalitarian has a duty to obey these laws. If so, violating the (spirit of) the laws to mitigate some of the unjust harms from which minorities suffer might not always be impermissible.⁸⁸ Indeed, as has been argued in the literature on (un)civil disobedience, it might even be that there is sometimes a duty to break the law as a way of resisting injustice.⁸⁹ Again, this deserves more attention, but it should at least address the immediate worry that simply because it goes against (the spirit of) the laws, it would be impermissible for the egalitarian to pursue indirect affirmative action in this way.

And now the third concern. Above, we considered the racist second-order discriminator who tries to game the system to secure racist effects by adopting a rule which does not single out a protected characteristic. If, as seems obvious, such racist second-order discrimination is impermissible, why isn’t egalitarian gamesmanship-style indirect affirmative action—which tries to game the system to secure egalitarian effects by adopting a facially neutral rule—impermissible as well?⁹⁰ The short answer: because the former is racist, and the latter is not. That they are relevantly different can be seen by deploying the following test; we may call it the “what if?” test.⁹¹ It is a counterfactual test which is meant to capture which role race plays in your action. Using it, you ask: if society had not been unjust—if we were under conditions of full compliance—would your view still provide a reason for treating some differently from others based on race? The strategic egalitarian can answer no to this question. If minority individuals did not suffer from injustices, there would not be a reason to pursue indirect affirmative action in favor of them, egalitarian gamesmanship-style.⁹² The racist second-order discriminator cannot answer no to this question. Their view is that Black people are morally inferior to White people: that the former’s interests intrinsically count for less than the latter’s interests. Their view would speak in favor of treating White people better than Black people because of their race, also, or perhaps especially, under conditions of full compliance. That they respond differently to the “what if?” test shows that there is a relevant difference between the strategic egalitarian and the racist in the way that race figures in their reasoning, and that, intuitively, what the latter does is objectionable in a way that what the former does is not precisely because of this. Indeed, the strategic egalitarian does not subscribe to a view on which one race is inferior to another. I do not mean to claim any novelty here: it is often argued, in the affirmative action literature, that affirmative action does not amount to reverse discrimination or racism.⁹³ Indeed, as Laborde points out: “Race- and sex-conscious policies of preferential treatment for members of dominated groups are not presumptively wrongful. They can contribute to reducing structural inequality, and they are not grounded in biased mental states and attitudes. Affirmative action in favor of African-Americans in the US, for example, is not an expression of anti-white racism.”⁹⁴ In other words: it is possible to maintain that racist second-order discrimination is impermissible, while maintaining that it may sometimes be permissible for the strategic egalitarian to pursue indirect affirmative action, egalitarian gamesmanship-style.

Of course, much more could be said here, both about the specific concerns I have addressed, and more generally about the permissibility of strategically pursuing indirect affirmative action. But as I said, my aim here was—in addition to showing that indirect affirmative action, because of its facial neutrality, is particularly strategically useful vis-á-vis direct affirmative action in a politically inegalitarian climate—to provide a preliminary case for the permissibility of sometimes pursuing indirect affirmative action as a form of egalitarian gamesmanship. These arguments, then, are meant as a starting point for further engagement with the distinction between direct and indirect affirmative action. Engagement which, despite the distinction’s significance, is presently largely lacking.