Structural Injustice, Exploitation, and Static Non-Ideal Theory

1. Introduction

Non-ideal theory is intended to be more directly relevant to the real world than its ideal counterpart. It is variously described as articulating, diagnosing, and proposing solutions for social problems (Anderson 2009); as placing methodological priority on rectifying injustice (Bohman 2012); and as more capable than ideal theory of recognizing historically important forms of oppression (Mills 2005). Yet non-ideal theory is not homogeneous. There are various ways theorists can account for, and engage with, how the actual diverges from the ideal. Moreover, non-ideal theory can be done better or worse, with greater or lesser success, or with more or less likelihood of generating positive change. And, as we will argue in this paper, some non-ideal theory can actually serve to obstruct attempts to make the world a better place, diagnose social problems, or rectify injustice.

More specifically, we defend two claims: First, while non-ideal theory is often thought of as transitional in nature–that is, theorizing how agents in structurally unjust circumstances should act so as to bring about a more just state of affairs–this transitional approach is not the only way to theorize obligations under these conditions. Theorists sometimes instead engage in what we call static non-ideal theory, which attempts to guide the actions of individuals while taking background structural injustice as a fixed point against which individuals can try to ethically interact. Second, contrary to the intentions of theorists, static non-ideal theory often serves ideological purposes by functioning to reinforce the (unjust) status quo.

We use a particular debate within normative theory–that about the moral status of exploitation–as an exemplar of static non-ideal theory and the mechanisms by which it plays this particular function. Our goal is neither to offer a theory of exploitation nor to defend a particular response to questions such as “when/why is exploitation wrongful?” or “should wrongful exploitation be permitted?”, but rather to make a methodological point that applies with equal force to theorists with opposing answers to these questions. Neither is our critique of static non-ideal theory intended to be limited to discussions of exploitation. We believe other debates in non-ideal theory are best understood as adjudicating complaints about static theorizing, and we return in the penultimate section to briefly illustrate this in a debate about effective altruism.

Our argument proceeds as follows. In section 2, we present a stylized account of how and why debates about exploitation often intentionally take background structural injustice as fixed, and we explain how this constitutes a non-transitional type of non-ideal theorizing, which we call static non-ideal theory. Sections 3 and 4 argue that holding injustice fixed in this context is problematic, because doing so functions to obscure important sources of background injustice and, therefore, important sources of obligations, resulting in theory which helps justify the status quo, thus playing an ideological role. Section 5 considers how, when, and why static non-ideal theory serves the ideological function it does in a world plagued by structural injustice, and provides some clues (but not a full story) for how to avoid being problematically static in our theorizing.

2. Exploitation Theory as Non-Ideal Theory

We engage with a relatively recent, popular brand of theorizing more consistent with the liberal political tradition than with Marxian thought. As will be demonstrated in this section, this line of theorizing, in one very important sense, deliberately sets aside facts about social structures and structural injustice in its determinations of what might constitute exploitation and what’s to be done about it. We focus on two contexts in which debates about exploitation are salient: the use of “sweatshop” labor in low- and middle-income countries (LMICs) by multinational enterprises (MNEs), and international clinical trials conducted in LMICs and sponsored by pharmaceutical companies or public research funders from wealthy countries.

In each context, there are at least two overlapping sets of questions theorists consider. The first set concerns what constitutes a fair versus an unfair or exploitative transaction, particularly under conditions of background injustice, and whether (and if so, why) a wealthier agent taking advantage of another party’s objectively bad situation constitutes a prima facie moral wrong. In the context of sweatshops, theorists ask whether sweatshop owners or MNEs which contract with them commit some moral wrong against LMIC workers or owe those workers better than the “sweatshop” wages and conditions they agree to work for, and if so how to understand that wrong. In the context of research, ethicists wonder whether (and why) researchers or research sponsors in some way wrong LMIC research participants or their communities, or whether participants or communities are owed benefits in excess of what they agree to receive in exchange for participating in or hosting research. In either case, worries that it may be wrong to take advantage of an agent’s bad situation confront both the fact that workers and research participants often benefit from these interactions in ways not otherwise available to them, and the corresponding intuition that it therefore cannot be wrong for such interactions to occur.

These questions intersect with a second set of questions, related to whether or how third parties should intervene to try to improve interactional outcomes for the worse-off party. Some worry that by trying to prevent exploitation, we may cause what Alan Wertheimer called “diversion effects,” ultimately preventing the relevant interactions from occurring at all, thereby preventing those who are worse off from improving their situations (Wertheimer 2015). The more benefits A must cede to B in order to interact with her, the less valuable the interaction is to A. If MNEs were required to ensure factory workers earn more, or research sponsors required to provide greater benefits or protections to research participants, the better off party’s costs of interacting would increase, reducing their self-interested incentive to transact at all.

For example, Matt Zwolinski (2007) has opposed both the creation of new labor standards in LMICs and greater enforcement of existing regulations because attempts to ameliorate working conditions in sweatshops (e.g. minimum wage or workplace safety regulations) increase the costs of employing sweatshop workers. These costs are passed on to MNEs who, in turn, might decide to do business elsewhere, where labor is more productive or less heavily regulated. Such greater (enforcement of existing) labor regulations might then “take away the option of sweatshop labor from the workers who would otherwise engage in it” (695). Similarly, in the case of international research, Wertheimer (2015) argued that ethicists “promulgating” standards demanding greater benefits for research participants–like requiring certain standards of care in comparator arms, or ensuring post-trial access to proven interventions–would undermine research sponsors’ incentives to conduct research in LMICs where disadvantaged populations could benefit from participation.

Importantly, even if these questions are analytically separable, such that we might identify some transactions as exploitative while believing they ought to be permitted to occur, our theoretically-informed claims that they are exploitative cannot help but influence actors in ways that could produce something like diversion effects. Perhaps we don’t expect corporate actors such as MNEs to respond directly to moral concerns about exploitation, but third parties like consumers might adopt a view of exploitation, and then influence MNEs to act on that view, like when university students organized in support of better labor conditions in manufacturing of university-licensed apparel (McKean 2020). To the extent that exploitation theorists are aware of these possibilities and want their theories of wrongful exploitation to be action-guiding, then, diversion effects can loom large even when focusing specifically on what constitutes wrongful exploitation (Gray 2020).

For example, Jeremy Snyder worries that a theory of exploitation that required paying LMIC workers a living wage to avoid being exploitative might lead to employers simply not hiring these workers, denying them beneficial opportunities. Snyder therefore proposes a theory in which employers can avoid exploitation by paying only as close to a living wage as is possible while maintaining an incentive to hire in LMICs (Snyder 2008). Snyder also worries that in the case of international research, “the fact that mutually beneficial and voluntary human subject trials do something to help the potentially exploited party speaks to the need to consider the consequences of setting the bar on these interactions too high” (Snyder 2012). Similarly, Michael Kates seeks a theory of sweatshop exploitation which will appropriately “balance the demands of fairness, on the one hand, with those of worker welfare, on the other,” noting that “one of the main objections to the very idea of a fair wage is that it makes sweatshop workers worse off by reducing the demand for their labor” (2019). In each case, worries about the impacts of a theory’s adoption inform the theory’s content.

Importantly, debates about these topics are adjudicated against a shared recognition that these interactions occur in a world rife with structural injustice, and that this injustice often creates the vulnerabilities that make exploitation possible. This recognition is what makes such theorizing non-ideal. But it is important to see just what kind of non-ideal theorizing exploitation theory is.

Gopal Sreenivasan (2012) gleans two distinctions between ideal and non-ideal theory from Rawls. One is that between full compliance and partial compliance theory. Rawls stipulates that parties in the original position can assume that the principles of justice, “whatever they are, will be strictly complied with and followed by everyone” (1971). At the same time, what justice demands in the absence of full compliance may differ from what it demands in the highly idealized world where everyone accepts and follows the rules.

Take the example of global poverty. If we believe the affluent have some obligation to contribute to poverty relief, then what each affluent individual should give might plausibly depend on what others can be assumed to be contributing. If everyone does their fair share, then each individual can be said to have the obligation only to contribute their part. But absent full compliance with such a norm, it is an open question whether individuals’ obligations remain the same or increase, for example, in proportion to either how much less others are giving, or how much more she can afford.¹

A second distinction is that between what Laura Valentini (2012) calls “end-state” theory and what we will call transitional theory. End-state theory concerns the features of the fully just (ideal) state, while transitional theory is meant to get us from the non-ideal here to a just (or more just) there.² Here, one relevant question is whether we need to identify what would constitute the fully just end state in order to make improvements to the actual, non-ideal world, or whether theorizing the end state may actually distract from or even obscure appropriate and feasible improvements over the status quo (see e.g.Mills 2005, Sen 2009, Simmons 2010, Gaus 2016).

Clearly, some theorizing is non-ideal in both senses, i.e. both partial compliance theory and transitional. This can also be gleaned from Rawls, who thought we needed an end-state theory that assumed full compliance to inform our non-ideal transitional theory, which itself assumed only partial compliance. But conceptually, partial compliance theory need not be transitional. A theorist might simply hold fixed the injustice resulting from partial compliance, for example, and then assign obligations which do not address this injustice. Theorizing justice might be left to others, in favor of focusing on obligations within (say) discrete interactions, which happen to occur against the backdrop of injustice. And the fulfillment of these obligations would not (nor be intended to) impact the fact of ongoing structural injustice, however that is defined.³ This suggests a complication to the above taxonomy, in that there actually exists a tripartite distinction between end-state theory, transitional theory, and what we will call static non-ideal theory–a kind of partial compliance, non-transitional theory.

Much of exploitation theory is static non-ideal theory: it recognizes the fact of background structural injustice but also takes it as a fixed point, and asks what a potential exploiter or third party ought to do or not do given injustice-based transactional (dis)advantages. So, for example, Zwolinski writes,

I am treating sweatshops as a somewhat isolated moral phenomenon. That is, I am asking what we should do about sweatshops, while holding most of the other conditions of the world (large inequalities of wealth among nations, severe poverty in the developing world, and a growing system of global capitalism) constant… Poverty, inequality, and economic development all need to be addressed. My paper seeks to tell us what we should do about sweatshops in the meantime (2007, emphasis added).

Kates explicitly describes the project of theorizing exploitation as one of formulating “non-ideal moral principles” for a world lacking full compliance (2019). He explicitly separates questions of what workers need or are otherwise entitled to as a matter of justice from questions of what they should be paid: “we are ultimately not interested in determining what sweatshop workers would be morally entitled to in a just world. Rather, we are interested in determining what a fair wage for sweatshop workers would be in an unjust world like our own.” And in discussing the ethics of LMIC research, Millum and Garnett note that, “One [question] concerns the unjust structural features of society, and what we are to do about them… The second… concerns what individual actors are to do in the meantime” (2020, emphasis added).

Millum (2016) explains this orientation: plausibly, agents have obligations to “work towards justice” by reforming unjust institutions–especially agents who benefit from injustice.⁴ But first, individuals–as individuals–simply lack the ability to do much about injustice. And second, the grounds of obligations to reform unjust institutions are importantly independent of potentially exploitative interactions. These grounds could be a general duty of beneficence, but obligations of beneficence are independent of specific interactions. Nor could considerations of beneficence require these parties to interact in the first place. Alternatively, obligations to reform unjust institutions could be grounded in participation in an unjust system, with attempted reform a kind of restitution. But this obligation “would not be related to any interactions between the beneficiaries and victims of injustice,” and “would stand even if [interactional parties] were otherwise isolated from one another” (Millum 2016; for similar reasoning, see Zwolinski 2012). So while it might be possible to discharge the attendant justice-based responsibilities via the terms of one’s interactions with parties from LMICs, such interactions themselves are neither the cause of, nor the only opportunity to discharge, those obligations. Given this, “such obligations do not… seem likely to affect the ethics of transactions against a background of injustice” (Millum 2016). Thus, we are to conclude, we can theorize about the ethics of individual transactions occurring in contexts of background injustice without that injustice impinging on the ethics of those transactions. We have an obligation to address structural injustice, sure, but in the meantime we want to know what obligations exist in this transaction.

Many exploitation theorists, then, regardless of their orientation towards the wrongfulness of exploitation or whether it should be permitted, recognize that many potentially exploitative interactions occur against a background of structural injustice, but adopt a division of theoretical labor:

1. One level concerns what must be done to mitigate or end structural injustice. Theorizing here seems to require answers to difficult questions like What is a just, or more just, basic structure?, and How can individuals best act to achieve greater social justice, given that progress plausibly requires collective action?

2. The other level concerns what agents should do in (potential) interactions of a specific kind, given the world as (unjust as) it is.

In the next two sections, we make the case that much of the project in (2) is conducted in a way that is both overly isolated from the project in (1), and inconsistent with the purpose of the project in (1). It is overly isolated from the project in (1) in that it makes assumptions that are incompatible with a serious recognition of the facts of structural injustice. And it’s inconsistent with the project in (1) because if those assumptions were broadly accepted, this would function to protect and reproduce the (unjust) status quo: A myopic focus on the obligations of specific, identifiable parties to interactions in isolation from the actions and conditions that circumscribe those interactions leads to a failure to recognize the role of powerful parties in creating and sustaining the unjust background conditions in which interactions occur. This results in obfuscation of important sources of obligations, such that theorizing that embraces this division of labor provides a kind of theoretical cover to those with the most resources and power, enabling them to continue to both maintain and benefit from those same unjust background conditions.

3. Framing Assumptions and the Division of Labor

In this section we show how within the exploitation literature, the division of labor described above leaves the project in (2) overly isolated from the project in (1). We argue that the prevalence of certain framing assumptions–that the transactions in question are consensual and mutually beneficial–obscures the ways in which entities like MNEs and research sponsors leverage their power on an ongoing basis to help shape the conditions in which transactions occur.⁵ This, in turn, obscures the ongoing interplay between structural injustice and the ethics of individual interactions.

To call the interactions in question “consensual” is to say that however desperate their situation, potential exploitees can at least give morally transformative consent to work or participate in research. To call these interactions “mutually beneficial” is to say that however little such workers or research participants receive for their contributions, by transacting they receive greater benefits than they would otherwise.

These assumptions are shared across a wide spectrum of exploitation theorists, and the consensual and mutually beneficial nature of the interactions is introduced as a kind of stumbling block for claims of wrongful exploitation: Given that the relevant workers and research subjects consent to and benefit from these interactions, how could these transactions be wrongly exploitative? Or, even if we do label them wrongfully exploitative, why would we consider taking actions (e.g. regulation, boycott) to prevent them, when workers and research subjects agree to interact and are made better off by doing so–particularly if such actions might deprive exploited parties of the ability to improve their circumstances?

These two assumptions are at the surface, but considerably more lurks below. Of particular importance are the assumptions that (i) no party is coerced into transacting (at least not by the other party); (ii) no party is morally required to transact; and (iii) the parties’ pre-transaction holdings constitute the relevant baseline by which to judge whether the transaction is beneficial.

Starting with (i), consent straightforwardly excludes coercion. (ii) may seem less obvious, but is related to the reasoning, given in section 2, for the division of theoretical labor under discussion. Here we borrow Gray’s (2020) terminology and refer to this assumption as no prior obligation: the MNE or research sponsor is under no special obligation to transact with the worker or the research participant to their benefit; if they choose not to interact, this would not violate some claim of the worker or research participant to benefits. No prior obligation is consistent with holding that the worker or research participant has been wronged in various ways, but not by the MNE or research sponsor in question; for example, by other agents or unjust social structures. It is also consistent with holding that MNEs or research sponsors have wide duties of assistance that could be met by helping people similarly situated to the worker or the research participant. But such duties could alternatively be discharged in various other ways, so the worker or research participant has no claim to the benefits to be gained by interacting with the MNE or research sponsor; and very importantly, the MNE and research sponsor are free (i.e. morally permitted) to not interact with the worker or research participant at all.⁶ Thus the struggle to explain what could be morally problematic about such interactions when they do occur. Point (iii) is not unrelated to (ii): we want to know if the terms A proposes to B are unfairly exploitative or whether outside parties should prevent the proposed transactions from happening. In making these determinations, theorists emphasize that by transacting on these terms, each party benefits compared to their status quo baseline.⁷

However, we must also consider that A may have acted so as to compromise the informativeness of this baseline. Sometimes, this can occur in ways that put pressure on the assumptions of consent and no prior obligation. Consider a case in which A has directly wronged B. Perhaps A stole something from B which B needs, then proposes to sell B some of that thing at the competitive market price without taking any steps to rectify the earlier wrong. Relative to their status quo baselines, each party benefits. Yet, despite this and whatever fairness we may or may not ascribe to A’s proposed terms under other circumstances, the way A has manipulated circumstances in his favor shows that the status quo isn’t always the relevant baseline: B is owed more than what would normally be considered fair, and failure to offer more may even constitute coercion if A owes B redress. Furthermore, knowledge of A’s prior actions is plausibly relevant to whether outside parties should intervene, since (for example) regularly permitting such transactions might incentivize further theft. Thus even given agreement on what constitutes exploitation, consent, status quo holdings, and potential relative benefit via interaction do not sufficiently complete the moral picture when one party engages in what we will call stage setting. We understand stage setting as leveraging (or, here, doing) injustice to rig the background against which interactions take place in order to strengthen or maintain their superior bargaining position relative to potential interactional partners.

Perhaps A directly wronging B is analogous to sweatshop labor and international research, but the case for this is not entirely straightforward. However, less direct forms of stage setting also put pressure on the relevance of the status quo. Suppose C, working in coordination with others, e.g. lawmakers and others similarly situated to C, has contributed to circumstances in which D is unlikely to be able to achieve better terms in filling some need than what C offers. Perhaps C has wielded influence to obstruct regulations or repeal protections for D that would have enabled D to effectively bargain for more. C’s stage-setting actions are more nuanced than A’s. Yet, because C has played its role in successfully preventing D from achieving more favorable terms, here too we cannot simply take improvements over the parties’ status quo baselines as the full story. Like A, C has worked to rig the game in its favor, and this is relevant to both the moral status of the proposed terms of interaction as well as obligations of third parties who might intervene: e.g. refusing to push for further (re-)regulation might allow the Cs of the world to turn profits gained from such rigged interactions into more lobbying power, which could then be used to further shape the regulatory environment.

Moreover–and crucially–C’s past actions are relevant moral features even if they are not the causal origin of D’s need. For instance, D’s need may have resulted from a long, complex history of race-, gender-, or class-based injustice in which C played no role, and no obvious causal connection can be made between C’s current holdings and the earlier injustice D has suffered. Still, C’s stage setting makes it insufficient to merely note that all parties stand to benefit relative to their status quo holdings.

Yet discussions of exploitation that adopt this framing generally proceed as if there is no analogous stage setting going on, so that the interactional background, and the relative bargaining positions it informs, are exogenous to the interactions themselves. While background injustice may be considered relevant for determining the distribution of moral obligations, it’s taken as a given or fixed point. And by theorizing as if exploiters play no active role in stage setting, the discussion is naturally shoehorned into a kind of faceoff between someone like the moral libertarian–who says consent is sufficient to guarantee the moral acceptability of a transaction–and those who argue for more substantive constraints on acceptability, e.g. a substantively fair division of the social surplus created.

But MNEs and research sponsors do in fact engage in the kind of stage setting described above: they wield their wealth and power to help shape the transactional background to their advantage, thereby calling into question the relevance of the status quo baseline. Corporate entities leverage their clout in favor of policy regimes that cement their superior bargaining positions and undermine the likelihood of preferable alternatives for those with whom they interact. This structuration can be active, including lobbying activities and participation in policy-making and agenda-setting, or passive, by way of the pressure that investment decisions place on states. Importantly, this kind of structuration is ongoing and continuous, occurring simultaneously with the kinds of transactions in question, further cementing corporate actors’ bargaining positions for future advantage.

For example, US-based multinationals lobby the US Trade Representative to prevent review of preferred trading status for countries in clear violation of international labor standards and threaten to pull investments from LMICs should those countries attempt to improve labor protections (Compa and Vogt 2000). The latter occurred when the American Chamber of Commerce in Shanghai and the US-China Business Council warned that proposed reforms in China’s Contract Labor Law would be “prohibitively expensive” and would damage foreign investment (McDonald 2007), leading to revisions in the final draft of the law that lessened protections for workers (Jing 2008). In an example of active participation in policy-making, European corporations, via the European Roundtable of Industrialists, exerted significant influence on the drafting of the Maastricht Treaty which established the EU, and again during the EU expansion into Eastern Europe from 1993–2004. The accession of Eastern European states into the Union was predicated on a series of liberalization and privatization reforms. These reforms ultimately led to thorough labor deregulation in the Visegrad countries of Poland, Slovakia, Czech Republic, and Hungary. In the Baltic states, the decimation of labor protections was so extreme that trade union density and collective bargaining is among the lowest in Europe (Cox 2019).

Even absent lobbying or active participation in policy-making, states must curry MNEs’ favor on a continuous basis, with the ever-present knowledge that investment decisions are “sensitive to political signals” such as changes in tax regimes and regulations (Woll 2019). LMICs construct export processing zones for the purpose of attracting MNE investment, often intending that normal labor protections not be enforced in these areas, leaving workers vulnerable to union- and strike-busting, excessive work hours, and unsafe conditions (McKean 2020). Clearly, states bear responsibility for enacting policies which jeopardize worker protections. But MNEs set a pattern of expectations via implied or explicit threats of capital flight which drive regulatory competition to attract or maintain investment (Woll 2019), and these activities function to actively construct the status quo for transactions and thus the baseline for what is considered consensual and mutually beneficial. Yet if an MNE works actively to create or sustain conditions in which their offer of employment cannot be reasonably refused, it is not so puzzling why the mere presence of consent and status-quo-benefit seem insufficient to guarantee non-exploitation.

International research sponsors engage in similar activities. For instance, the pharmaceutical industry–a primary funder of international research–was a major proponent of the World Trade Organization’s Trade-Related Aspects of Intellectual Property Rights (TRIPS) agreement. Prior to TRIPS, many LMICs limited patents on medicines–reducing patent terms, allowing patents only on manufacturing processes, or excluding pharmaceutical patents altogether (Hoen, Berger et al. 2011). In contrast, TRIPS requires all WTO members to recognize strong patents on medical technology, including pharmaceutical interventions, undermining the previously successful generics manufacturing industry in LMICs. The lack of generic competition drove increases in the costs of many essential medicines (Tenni, Moir et al. 2022), resulting in reduced access to medicines, particularly in middle-income states, which host the majority of LMIC research (Jung and Kown 2015). Since alongside altruism, access to healthcare is among the primary motivations for LMIC research participants (Browne, Rees et al. 2019), we can trace a line from pharma lobbying to the ability to recruit LMIC participants into studies for fewer benefits than might otherwise be necessary. In using their significant power to promote and protect strong intellectual property rules governing medicines, the pharmaceutical industry shapes the background context of international research, as comprised in no small part by limited access to medicines, which informs the kinds of agreements that prospective research participants and the communities that host research are willing to accept. And they do so on an ongoing basis, for example by leveraging economic threats against WTO member states to discourage use of compulsory licenses in those circumstances permitted by TRIPS provisions (Schüklenk and Ashcroft 2002).

In each of these examples, powerful entities who enjoy a bargaining advantage over individuals and communities in LMICs bear significant responsibility for structuring the background against which their “mutually advantageous, consensual” interactions take place. What is gained by recognizing the existence of stage setting in these contexts? First, it highlights a crucial disanalogy between these large-scale exploiters and the small-scale exploiters–e.g. individual actors and other relatively small actors–often discussed in the exploitation literature.⁸ In the latter, it is plausible to assume that all actors are causally inert with regards to the background against which they interact: this background is merely a given, and happens to provide one party a threat advantage over the other. A tow-truck driver, for example, is likely not responsible for the fact that their customer (and potential exploitee) needs urgent assistance. Exploitation occurs here, then, when the stronger party wrongly leverages this given advantage. But unlike small-scale exploiters, MNEs and major research sponsors do not merely happen upon some vulnerability and face a question about whether to take advantage of it; nor do they merely seek out pre-existing bargaining advantages. Crucially, they also actively work to manufacture and sustain the background conditions that enable their further extraction of value from LMIC populations.

Clearly MNEs and research sponsors cannot achieve this feat by themselves: they must collude with other powerful entities, most obviously governments and perhaps international governance bodies as well, all of whom share responsibility for the transactional context. But that MNEs and research sponsors have played their (very important) roles in all this is surely relevant for understanding both what obligations they’re now under in their continued interactions with LMIC populations, and how we should understand our own obligations qua agents for potential intervention and change.

One might think this shows only that exploitation theory conducted on the basis of the division of labor discussed above has typically failed to take into consideration something that it should; not that it is inherently incapable of grappling with stage setting. Accepting the division of labor, one might identify two sets of potentially relevant variables for normative theorizing. Corresponding to the second part would be interactional variables, such as the moral force of each party’s consent, whether and how much each party benefits relative to some relevant baseline, the parties’ non-transaction costs, and whether one party owes the other some amount above what the other would consent to transact for. Corresponding to the first part of the division of labor would be structural or background variables: considerations relevant for assessing the justice of the background structure, against and within which transactions happen to take place. These variables include, among others, the background distribution of resources, the distribution and exercise of power, and potentially ethically relevant historical interactions.

Upon recognizing the fact of stage setting, one might concede that the values of interactional variables cannot be correctly assigned without assessing the validity of the values assigned to the background variables. And this seems plausible: what looked like consent might turn out to have been coercion, what looked like a substantial benefit might be less than would have otherwise been expected. Yet one might simultaneously insist that interactional variables can nevertheless be sufficiently informed by the (correct) background variables so as to enable us to carry on with the second project identified above. In other words, one might hold that while interactional variables are not independent of structural ones, once the former are correctly informed by the latter, interactional ethics can then proceed independently of theorizing the nature and means of achieving structural justice.

But this brings us to the second, and more fundamental upshot of recognizing stage setting in these contexts, and it comes to light by pointing out a further disanalogy between the types of small-scale interactions exploitation theorists often work with and the large-scale type of interactions we’ve concerned ourselves with here. The former tend to have a discrete, one-and-done character, whereas sweatshop labor and international clinical research involve continuous, mutually-reinforcing relationships between background and interactional variables. Stage-setting activities are not incidental to the fact of current and future transactions: rather, past actions were taken in many cases with at least partial intent of preventing other potential parties from having a stronger bargaining position in later interactions. Similarly, parties in bargaining positions weakened by stage-setting activities are precluded from demanding more and are thus unable to improve their condition sufficiently to demand more in the future. The resulting transactional outcomes contribute to future distributional facts, their future weak bargaining positions, and the corresponding greater ability of powerful actors to continue their stage-setting activities. There is constant interplay between the two sets of variables. And as long as this mutually-reinforcing process is ongoing, there can be no neat theoretical division of labor that does not function to obscure centrally important aspects of the relations between the parties.

We can now more clearly specify how exploitation theory often leaves the project in (2) (theorizing the ethics of individual interactions, like sweatshop employment or LMIC research) overly isolated from the project in (1) (theorizing structural injustice and what to do about it). Mainstream theorizing about what is or isn’t exploitative in these interactions, or whether these transactions, if they are exploitative, should take place, takes no account of powerful parties’ continuous involvement in structuring the conditions under which these transactions occur, while this structuring is at least part of the reason the terms sufficient to garner consent are both so paltry for recipients, while also so beneficial compared to non-transaction. In insisting on the theoretical division of labor, this important interaction between the sources of structural injustice and the backdrop for individual transactions is lost.

4. Exploitation and Ideology

Above, we argued that the division of labor some theorists prefer leaves the project of theorizing interactional obligations overly isolated from theorizing about structural injustice. In this section, we make the further claim that the presumptions baked into this prevalent narrative are also inconsistent with the project of addressing that injustice. Ultimately, we are making a claim about ideology, and our critique has both an epistemic and a functional component (Shelby 2003). The epistemic complaint is comprised primarily of the observations of section 2: debates about exploitation are framed in a manner that distorts important features of the interactions in question. Specifically, this framing obscures the various ways in which the contexts of exploitative interactions attenuate the moral weight of consent and the validity of the pre-transaction baseline. The functional complaint, which we lay out in this section, is that this underlying framing plays the academic and social function of legitimizing ongoing modes of interaction that entrench and reproduce the same unjust power asymmetries that structure those interactions to begin with. When exploitative interactions are theorized without attention to the interaction between the nature and sources of background injustice, the unjust status quo is reinforced.⁹

First and most importantly, theorizing this way takes the most effective tools for recognizing and redressing the wrongs that make ongoing exploitation not merely possible, but apparently “beneficial,” out of play before debate even begins. Exploitation is theorized entirely as advantage taking, with no room for or engagement with the extent to which it is simultaneously, and often primarily, advantage creating and maintaining. Without attention to the ways large-scale exploiters continuously create and maintain their advantage and the corresponding positions of disadvantage of those with whom they interact, we cannot conceptualize why obligations to engage on better terms might arise. The result is a limited set of analytical tools that leaves theorists across the range of substantive views on exploitation scrambling to ensure that the conditions for ongoing exploitation are maintained for fear of depriving needy exploitees of the meagre benefits on offer. This serves to both let those with disproportionate power and resources off the hook should they choose not to interact with disadvantaged populations, and also make us question why, if they do choose to so interact, they should have to offer more than the minimum needed to achieve consent. And in so doing, it also legitimizes the further entrenchment of the very conditions that facilitate this exploitation, enabling further accumulation of wealth and power by those who already hold more than their just share, and ensuring that those who are exploitable now remain so into the future (Wenner 2015).

Second, and following from that, in presuming a context of consent and mutually beneficial transaction, this theorizing normalizes assumptions about the location of the relevant normative baseline. As we have shown, the idea that it’s puzzling what could be wrong about these interactions–since they can be presented as consensual and mutually beneficial–requires one to first ignore the very real fact of stage setting. The more myopic our theorizing is here, the more legitimacy is granted to the status quo distribution, and the more difficult it is to see why powerful entities aren’t allowed to simply dispose of their resources however they see fit.

Third, and relatedly, this legitimizing of the baseline results in a serious, self-reproducing methodological error. In many cases, engaging in the relevant debates at all requires theorists to adopt the problematic framing and division of theoretical labor we’ve identified, making it even more difficult to challenge the legitimacy of status-quo distributions of power and resources, resulting in theory poorly equipped to provide traction on the very problem it purports to target.

Fourth, and finally, the framing of exploitative interactions as mutually beneficial and consensual elides the fact that in some instances, consent that is sufficient to alter the agent-relative moral permissibility of interacting with someone may nevertheless not amount to fully autonomous consent when an agent’s option set is significantly constrained. Granting that individuals living in impoverished conditions can give morally transformative consent to employment or research participation, it is nevertheless the case that continued deprivation undermines full agency, in the sense that those in better conditions are more empowered to reject inadequate offers. Obfuscation of this fact serves to both sanction and sustain conditions under which agency cannot be more fully developed (Preiss 2014).¹⁰ In treating consent as the bulwark of legitimacy, the shared framing discussed above allows for the status quo to perpetuate itself via consensual interactions that contribute nothing to the future ability of those positioned as exploitees to secure sufficient agency to interact more freely in the future. Instead, it introduces noxious incentives to MNEs and research sponsors to continue their efforts at stage-setting so as to maintain and deepen their significant threat advantage (Pogge 2008). More to our point, it insulates exploiters from justice-based demands to offer interaction on more favorable terms.

5. Static Non-Ideal Theory

We can now posit a tripartite division of political theory: (1) ideal theory as end-state, full compliance theory, modeling the perfectly just society; (2) transitional non-ideal theory, as partial compliance theory focused on achieving a (more) just society, either in conjunction with an end-state theory or not; and (3) what can be called static non-ideal theory, as partial compliance theory which is non-transitional because it holds as fixed the facts of structural injustice, rather than theorizing how to address those facts. Importantly, holding background injustice fixed, we might consider that background injustice more or less relevant to ethical theorizing about the obligations of particular agents. In at least some important cases of static non-ideal theorizing, agents’ obligations are theorized as if the facts of structural injustice need not inform them, often as if these parties are causally inert with regard to those facts even when they are not. It is this kind of abstraction from background injustice that often serves to reinforce and perpetuate an unjust status quo, thereby serving ideological ends.

As Young noted, structural injustice is not some naturally occurring misfortune (Young 2006). It is the product of individual, intentional choices, even if the intention is not always to bring about or reinforce injustice (though in some cases it clearly is). Under such conditions, how agents act is in some sense both a product of the underlying structure (the system of incentives in place and the option sets available) and a potential opportunity to either change or maintain this same structure. And some agents within social structures have more ability than others to impact those structures, including the ability to effectively influence the background distribution of power and resources and the rules for how such are distributed to begin with.

Recognition of these differences in power among agents must be multidimensional. For example, the ability to set the terms of an interaction due to one’s threat advantage is one kind of power, but so is the structural power of being able to resist changes (e.g. in regulations) that would threaten that advantage. Success in the former (e.g. paying low wages) ensures greater resources to put to use in the latter (e.g. lobbying); and the latter clearly reinforces the former. More generally, when different types of power are mutually reinforcing in this way, we cannot theorize the ethics of wielding power in one dimension in complete abstraction from the other(s). And mutually reinforcing dimensions of power are what we should expect when theorizing ethics within unjust structures.¹¹

This suggests that theorizing the ethics of differently situated groups within structural injustice cannot be effective unless it acknowledges and engages with the hegemony of dominant groups–including their influence over the conceptual tools we use to understand and assess our world–and the extent to which current structural injustice is largely the result of how power is and has been intentionally wielded. This in turn requires engaging with the causes and sources of injustice, asking how these can be remediated, and considering how the assignment of obligations to dominant parties can be leveraged to effectively undermine distributions of power which allow some to wield mutually reinforcing forms of it in ways that cement their positions as dominators.

One potential worry for our view is that while some entities have power to help create and reproduce structures which insulate their advantages, other entities have extremely little or no power in this regard. Returning to sweatshop exploitation, for example, we might grant that MNEs have this structural power, but acknowledge that the owners and managers of sweatshops seemingly do not, as their practices are largely determined by the terms of contracts dictated by corporations. Their actions do reinforce structural injustice and thereby further exploitation, but such actors have far more limited option sets and less ability to impact unjust structures. We might still want to ask what obligations these actors have–perhaps we do, in fact, still need to figure out what to do about sweatshops “in the meantime.” And it might be argued that static non-ideal theory is appropriate and even valuable for evaluating the actions of these far less powerful actors.¹²

However, once we acknowledge that some actors are far more capable, and others far less, of creating and sustaining unjust structures, we might plausibly question the value of a heavy theoretical focus on the ethical analysis of those less powerful actors. In fact, and as we’ve argued, a widespread focus on such actors lends itself naturally to the kind of ideological function we’ve highlighted. Moreover, once it’s admitted that the actions of less powerful actors do play a role in sustaining injustice, this fact must at least enter into our considerations of what they ought to do. If it turns out such actors have no real choice but to continue reinforcing ongoing injustice, then ethical theorizing here (to the extent it can be done) might have results looking similar to those produced under the division of labor of static non-ideal theory. But that is importantly different from actually embracing that division of labor. In static non-ideal theory, the role(s) of actors in reproducing unjust structures is set aside. In contrast, in transitional theory this role is always a source of moral obligation, though one that is both contextualized to an agent’s structural position and may be weighed against other considerations (e.g. how much sacrifice can be morally required in obligations to address injustice, or how much loss of potential benefits to victims of injustice is acceptable).

Our critique of certain static non-ideal theorizing mirrors in important ways Charles Mills’s ideological critique of ideal theory. Ideal theory, Mills argues, assumes away actual histories of injustice and conceptualizes individuals as self-sufficient agents engaging in free and fair transactions; but this ignores the ongoing impacts of injustice and important sources of obligation, thereby functioning to preserve illicit group privilege and depriving us of the perspective necessary to theorize how best to promote justice (Mills 2005). According to this view, non-ideal theory should be better situated to address injustice due in part to both its acknowledgment of existing injustices, and its orientation towards rectifying them–in other words, non-ideal theory is taken to be transitional. Whether one finds Mills’s concerns persuasive against the value of ideal theory, the kinds of worries he levies are certainly applicable to much static non-ideal theory: the use of a theoretical framework that abstracts away from how injustice is perpetuated works to mask how the fact of domination structures our interactions with one another, and in doing so functions to reinforce and perpetuate status quo injustice. Non-ideal theory can be transitional, yes, but so too can it be static.

Finally, it is worth highlighting that the distinction between static and transitional non-ideal theory may help to bring clarity to other debates in political theory. For example, one prominent criticism of effective altruism is the so-called “institutional critique.” The general worry of the critique is that effective altruists commonly take existing social and political institutions as fixed, asking “how can this individual effect the most change?” (See, e.g., Gabriel 2017). This focus on individual contributions (so the criticism goes) tends to obscure both the sources of, and most likely collective solutions to, unjust social structures by failing to consider how institutions might be reformed to better promote justice (Dietz 2019). This results in what Amia Srinivasan has characterized as a rather “conservative” movement, allowing those who benefit from the status quo to comfort themselves that they’re doing the most good they can, while really leaving everything “just as it is” (Srinivasan 2015).

Of course, effective altruists may have excellent responses to this critique, perhaps most obviously that it’s not a critique of effective altruism at all, but of a misguided understanding of how to do the most good: to the extent that the critique is correct, effective altruists should recommend those efforts necessary to reform unjust institutions, rather than taking an overly-individualistic approach which leaves structural reality “just as it is.”

However, an alternative interpretation of the institutional critique reads it not as offering disagreement about where and how individuals can do the most good, but rather as expressing a concern similar to those we’ve outlined about exploitation theorizing. Both exploitation theory and overly-individualistic effective altruism demonstrate (a) a heavy theoretical focus on (b) the ways individuals can act so as to (c) improve (or not worsen) the lives of the worst off, while (d) working against the backdrop of grave structural injustice. In each case, while theorists do, or at least can, recognize the need to meaningfully address that background injustice, the reality that individuals generally cannot effect structural change on their own often results in theorizing that (a*) takes theoretical attention away from (b*) the kind of collective action necessary to (c*) improve the lives of the global poor in a more durable way by (d*) addressing the root (structural) causes of injustice. In each case, the problem is described such that global structural injustice is taken as a fixed point not meaningfully impacted by individual actions, directing theoretical attention instead to how individual actions can best be leveraged to improve welfare within the context of that injustice. As in exploitation theory, a complaint of static theorizing here is not a complaint about individual actions, e.g. those who respond to effective altruism’s reasoning by seeking larger incomes to donate to poverty relief. Rather, it is a meta-theoretical worry about where and how we choose to focus theoretical attention, and the ways that attention can obscure, distract from, and perhaps even help sustain the sources of injustice that are most pervasive in their effects.

Conclusion

Our aim has been twofold. First, we sought to carve out a space in the taxonomy of political theorizing that allows us to identify and call out non-ideal theorizing that functions to reinforce the non-ideal circumstances it responds to. And second–and to that end–we sought to demonstrate how in the case of exploitation theorizing, this ideological function is often served by an ill-advised division between interactional and structural foci within normative theorizing. Divided in this way, normative theory doesn’t just fail in its task of recognizing the true nature of the problems it’s meant to address. Worse, this failure results in hiding both responsibility for these problems and obligations to effectively address them, thus allowing wrongs to propagate on the hope (of increasingly strained sincerity) that one day–after the meantime–we’ll finally work out that other part of our theory and actually leverage it to do something about injustice. In other words, by holding injustice constant, static theory doesn’t just fail to call out and address injustice–it is frequently an unwitting agent of injustice. Whatever the context, where structural injustice or background power disparities condition social reality, theorizing about agents’ obligations can engage with and respond to those background conditions to a greater or lesser extent. And any time theorizing asks what agents ought to do against an allegedly immutable backdrop of structural injustice, we must ask whether this ethics of powerlessness is in fact ultimately an ethics for the powerful.

We have not offered a full account of what effective transitional non-ideal theorizing might look like. However we hope that other theorists will find this methodological intervention compelling, and join in the development of more productive theorizing about the most serious problems of our day. The structural injustices that theorists seek to define, debate, or set aside have very real implications for huge swaths of the global population, and to the extent that our theorizing serves to prop up or give cover to those injustices, we become complicit in them. We must collectively recognize that we can either theorize what to do to fight injustice–using all of the tools available to us–or we can continue spinning our theoretical wheels in support of the status quo.

Acknowledgments

The authors would like to thank Zoe Schneider, participants in the 2021 Center for Ethics & Policy Workshop on Political Philosophy & Bioethics and the 2021 MANCEPT workshop Towards a Unified Theory of Exploitation, as well as audiences at the 2020 American Society for Bioethics and Humanities and the 2021 Association for Practical and Professional Ethics for their helpful feedback on previous iterations of this manuscript.