On Trade and Exploitation

1 Introduction

Risse and Wollner’s (henceforth RW) On Trade Justice: A Philosophical Plea for a New Global Deal takes us on a theoretical tour de force that ranges from the grounds of justice to the nature of exploitation and, relatedly, power, fairness, reciprocity, just wages, and much else. The book also provides a wide range of policy prescriptions about, among other things, production offshoring, export subsidies, how to reform the current WTO system, and why we should normatively reject the lurking threat of mega-regionalism. Given the breadth of these concerns, it would be impossible to fully engage with the book as a whole.

In what follows, I shall concentrate on the idea of exploitation. In Section 2, I offer a sketch of RW’s account of exploitation. In Section 3, I tackle the question of whether exploitation would exist in a perfectly just society and conclude that RW have not offered us good reasons to think that it would. Similarly, in Section 4, I argue that RW have not shown that the WTO system is exploitative. To claim that members of a given institution are exploited we need a clear sense of what they ought to receive from it. But, RW do not provide a clear benchmark. In addition, I believe such a benchmark would not be an easy one to provide for the simple reason that determining what the (net) gains from trade are and how they are apportioned within the WTO is extremely difficult. In the fifth and final section, I address the issue of fair working conditions and wages. There, I contend that RW have not offered us a convincing reply to what I call the ‘market pressure argument’ against firms’ ability to improve working conditions, but that providing it is essential. I then discuss their proposal to assess the fairness of wages and raise a number of objections both on the cost side and the contribution side.

Before moving on, allow me one caveat. None of what I say here should be taken to imply that exploitation never takes place in the global economy, that the WTO is a perfectly just institution, that no person is exposed to horrible working conditions (or that it is a good thing when they in fact are), or even, more abstractly, that whatever outcome is produced by market forces is ipso facto fair. My qualms with RW’s work are, in the main, about their philosophical arguments, not their general orientation.

2 Risse and Wollner on Exploitation

Fair trade is, according to RW, intimately connected to exploitation. To trade fairly, in short, is to trade in the absence of exploitation. More specifically:

the principle of justice associated with trade as a ground, the principle regulating the internal workings of trade, requires the absence of exploitation: that is, the distribution of gains from global trade is just only if these gains have been obtained without exploitation. (RW 2019, 78, emphasis in original)

Is it plausible to portray all injustices that take place within the global trading regime as related to exploitation? After all, some have depicted the basic moral flaw in global trade as a case of stolen goods (Wenar 2017). Others have insisted that trade is part of a global economic order that harms the poor (Pogge 2002). The list could go on (see Barry and Wisor 2014; James 2012 for an overview). Yet, to their credit, RW (2019, 79) avoid this kind of reductionist picture. At the core of their strategy is the distinction between moral wrongs as violations of principles that arise in the context of trading and wrongs that arise from trading, the former being a much wider category than the latter. And it is the latter that they define as the absence of exploitation.

What, then, is exploitation? RW (2019, 78, 80) start their account of exploitation by offering the broadest possible view of the concept. In the generic sense, to exploit is to take unfair advantage. In addition, and this is far less obvious (though I think correct), exploitation is a sui generis moral wrong that cannot be captured simply by looking at the distributive outcomes of a transaction, nor solely at the way in which the transaction took place (RW 2019, 79). More formally:

Exploitation is a hybrid wrong with a combination of interaction and allocation at its heart. … we understand exploitation as a transfer T or distribution D between parties A and B, which arises as a consequence of interaction I, enabled by some ex-ante features F, violating some moral principle P, such that the overall moral defect cannot be readily reduced to a defect of either T, D, I or F. (RW 2019, 81, emphasis in original)

This formal definition of the concept, they claim, captures the structure of ordinary discourse about exploitation, and allows us to have a framework through which we can map the basic disagreements between different conceptions.

What is RW’s conception of exploitation? After a short but very incisive overview of the existing literature, RW (2019, 88) opt for what they call a general conception of exploitation as ‘unfairness through power’. The general account is defended as an ecumenical response to the persistent disagreement among different conceptions of exploitation and as intuitively appealing (RW 2019, 89). Following Vrousalis, they see exploitation as belonging to the ‘realm between coercion and unfairness’ (Vrousalis 2013, quoted in RW 2019, 89). Exploiters use power over exploitees (RW 2019, 90). Yet exploitation also (necessarily) involves the violation of some distributive norm; a violation that typically occurs ‘when in cooperative interactions the distribution of the surplus fails to satisfy all relevant claims’ (RW 2019, 91). Starting from this general conception, RW move on to provide a trade-specific one. Within the context of trade, we should think of exploitation as a ‘power-induced failure of reciprocity’ (RW 2019, 88).

RW’s trade-specific account of fairness claims is as follows:

Since trade is a particular type of structured cooperation for mutual gain, fairness requires all cooperation-relevant claims be satisfied proportionately. Within structured cooperation for mutual gain, the basis of claims is provision of benefits to co-operators. Actors have a claim of fairness to reciprocation against those whom they benefit and who voluntarily accept. Questions of fairness arise as questions of reciprocity. … There could be failures [of reciprocity] because actors bear disproportionately large costs of cooperation or because others receive disproportionate benefits. (RW 2019, 94)

Such failures of reciprocity, when exploitation occurs, are induced by power (RW 2019, 95). Power of some over others might arise, RW tell us, in different ways: from distributive or past injustices to mere accidents, a wide array circumstances generate bargaining differentials that allow its exercise. Power can also be wielded in different ways: through roles that assign decision-making authority, through force, or threats. The main examples that RW (2019, 95) provide, and to which we shall return to in the next parts of the essay, are wealthy states using their superior bargaining power to ‘uphold trade barriers while poorer countries are required to reduce them’, or corporations when ‘they benefit disproportionately at the expense of workers’.

Finally, RW (2019, 107) address the thorny topic of whether we should always condemn exploitation. In other words, are there circumstances in which ‘violations of trade justice can … be all-things-considered permissible?’ The question arises from the fact that exploitation, as it has been defined above, is a normal occurrence within the global economy, and, perhaps more importantly, seems to be inherent in the export-led growth model that has lifted millions out of poverty (RW 2019, 108). In this picture, it might be plausible to tolerate exploitation (see Powell and Zwolinsky 2012; for a criticism see Coakley and Kates 2013). The argument in favour of the latter conclusion can, according to RW, take two forms (2019, 108). Exploitation might be tolerable in order to create a more just world in the future (what they call the stepping stone argument). Or it might be tolerable in light of the benefits that it confers on its victims (what they call the price worth paying argument).

Both arguments need to be addressed in light of two distinct normative questions: (a) whether exploitation is necessary; and (b) whether it is permissible (RW 2019, 109). RW’s answer is that whether exploitation is necessary in the relevant sense depends on who deploys the argument and whether the proponents, individually or jointly, could avoid exploiting (RW 2019, 111). When it comes to permissibility, RW’s (2019, 112f.) argument becomes more complex, and basically requires additional (necessary) conditions. More specifically, exploitation will be permissible when it is necessary and: (a) the moral costs associated to exploitation are not (normatively speaking) too high; (b) the magnitude of the good to be achieved via exploitation and the probability of achieving said good are high enough; (c) the feedback effects between necessity and permissibility are not pernicious (e.g. they do not allow something morally wrong to become necessary because it is deemed to be permissible); and (d) instances of exploitation can be plausibly justified to individuals understood as separate persons rather than simply as a result of consequentialist considerations.

3 Some Concerns About Exploitation in Ideal Theory

Later in the essay I shall have something to say about specific applications of RW’s account of trade justice. For now, I’ll concentrate on the relationship between trade justice and ideal theory. The basic question we can ask, with RW, is as follows: would we still be in need of an account of trade justice based on exploitation if we lived in a just world (here, see Walton 2019)?^[1] RW’s answer is ‘yes’, but I believe that the arguments they offer in support of their view are unpersuasive.

Allow me to explain why I wish to press RW on this particular point. One might be inclined to think that the issue I am tackling here is not a central one (alas, the world is far from perfectly just). Three reasons militate against that conclusion. First, the relationship between justice and non-ideal theory has been, of late, central to methodological debates within analytical political philosophy (see Sangiovanni 2007; Valentini 2012). Second, the distinction they use to support their claim (between discretionary and structural exploitation) and especially the examples they use to illustrate it, are important to the book’s project as a whole. For they purport to individuate a class of cases when a corporation, or a state, can exercise power in a domestic or global market economy. That the class of cases is remote (since it presupposes a fully just society and/or world) doesn’t make it any less (theoretically) interesting. Third, and most importantly, because the absence of exploitation in ideal theory would imply that trade justice is not a distinct ground of justice; trade justice would ‘arise from other grounds in non-ideal theory’ (RW 2019, 69).^[2]

Going back to the main question: would exploitation take place in a fully just world? To begin with, the question is indeterminate unless one specifies an account of justice. Following RW, let us assume that a just world is one in which every political society is internally ordered in accordance with some version of justice as fairness. In such a world, RW (2019, 69) tell us, there would still be markets, there would still be exploitation, and thus, trade justice should also be theorized in ideal theory. The distinction that they draw in support of their argument is the one between ‘structural’ and ‘discretionary’ exploitation (2019, 69). They define and illustrate the distinction in the following way:

One way in which a company might exploit is discretionary: this is behaviour which companies could stop without putting their own business at risk (e.g. abusive behaviour by management). But exploitation might also occur structurally. Were a company A to cease such exploitation, this would undermine its operations due to the conditions under which it operates (e.g. small profit margins). (RW 2019, 69)

RW concede that in a just world exploitation is largely absent. Yet, they maintain that there could still be discretionary exploitation. My sense is that the examples they use to illustrate the distinction between structural and discretionary exploitation do not support the claim that discretionary exploitation is a plausible occurrence in a just world. More broadly, even if discretionary exploitation is possible in a just world, it is unlikely to be result of the existence of markets.

My argument, in short, is as follows. If labour markets are perfectly (or close to perfectly) competitive, discretionary exploitation in the form of abuse (i.e. one of the two examples RW use to argue for the persistence of discretionary exploitation in a just society) would disappear. This is because specific firms would have no (rational) reason to abuse their employees and would be unable to exercise power over workers when those workers have identical outside options. What would happen if labour markets were not fully competitive? The answer is that, assuming society to be fully just, the state would intervene to correct market imperfections affecting such a sensitive domain and, in any case, would not allow workers to pay the cost of such imperfections while they last. What if the state was simply unable to do what I am claiming it would do? Then, my sense is that we would be facing problems connected to unfavourable conditions, and this would represent a violation of the assumption that we are dealing with a just society facing favourable conditions (see footnote 1).

Let me expand on some of the points I have touched upon. Consider what would happen to a company that tolerated abusive behaviour by management in the context of competitive labour markets. Doing so would be just about the most counterproductive thing management could do from its own point of view. For, if labour markets are competitive, then abusing your employees just means sending them, and the time and money you have invested in them, off to the competition. One way to think about this is to see abuse as something that diminishes workers’ compensation without, presumably, improving their productivity. If the labour market clears, and if we have a specific wage rate in equilibrium, no employer can diminish compensation without losing workers (see Borjas 1995/2008). Put differently, in a perfectly competitive market abused workers would have a very clear outside option: the same wage without the abuse. More broadly, threatening a worker by saying ‘it is my way or the highway’ when the ‘highway’ is better than ‘my way’ is not much of a threat and thus is difficult to represent as an exercise of power.

What if, more realistically (or perhaps less fantastically), we only assumed a just society coupled with imperfectly competitive labour markets. Here, think of search costs for both employers and employees, monopsony power by employers in specific sectors, and short run macroeconomic fluctuations that do not allow the economy to operate at full capacity (resulting in unemployment). Would we have exploitation (discretionary or structural)? My sense is that the answer is still ‘no’. If we take the idea that we are in a just society seriously, we should also ask why market imperfections that affect the quality of workers’ alternatives to their present employment (or, even more markedly, their lack of employment) would be allowed to remain in place. Presumably, in a just society the state would adopt countercyclical macroeconomic policies, discourage or regulate monopsonies, and try to reduce search costs by developing institutional capacity to that effect. Even more importantly, in a just society the state would be compelled to provide a level of unemployment benefits allowing workers not to accept just about any job (or working conditions) for fear of destitution. And this is because one of the main purposes of embedding markets in the wider framework of a conception of distributive justice is precisely to improve workers’ outside options in the labour market, and do so to the extent that they should not feel compelled to accept working conditions that they deem to be unfair (for example, see Freeman 2013).

Lastly, what if, whether labour markets are perfectly competitive or not, prevailing social, economic and political conditions in a country are so dire that the state cannot perform the role that we might think it would plausibly perform in a just society? In that case, we would be leaving the realm of ideal theory. More specifically, the society in question would be facing unfavourable conditions, something that RW assume it would not in their description of a just world.^[3]

What about other instances of ‘ideal theory exploitation’ cited as examples by RW? The other main example RW (2019, 69) put forward is based on the idea that even in a world where each society is internally just there could still be large inequalities among societies determining vast differences in bargaining power between them. Once again, I am unconvinced. For exploitation to take place we not only need specific agents to enjoy the kind of power that would allow them to exploit, we also need them to have the desire to exercise that kind of power. Yet, in a just international society both of them would be in short supply. At the very least, this would be the case if we use Rawls’s view of domestic and international justice as a background picture (something that RW openly suggest that they are).

The finer points of Rawls’s work on international justice need not detain us here (see Wenar 2006, but also Reidy 2004, and Maffettone 2017). What seems relevant for our purposes is that a just liberal society is conceived of, by Rawls, as a well-ordered people. And a well-ordered people does not have the kind of interests that would make it plausible for it to exploit other well-ordered peoples by driving hard bargains. The idea of a liberal people is a normative conception and its foreign policy is an extension of the kind of agent a people is: its interests lie, principally, in the safety and security of its citizens and in its freedom and independence, not in the accumulation of riches or power. A people does not have a conception of the good to further, and so obtaining more resources beyond what is required to sustain just institutions is not part of its rational desires.

Furthermore, since we are in ideal theory, we are once again assuming favourable conditions, and that implies the absence of what Rawls calls burdened societies. Thus, not only can we conjecture that richer peoples would have no interest in using their superior bargaining power over other members of international society, we can also conjecture that poorer ones would have no compelling reasons to accept such bargains, or any bargain that they would deem to be exploitative.

Finally, let me address two objections to my criticisms of RW in this section.^[4] The first is that I am assuming too idealized a picture of a just society and of a just world. There is no need to assume that, in a just society (or world), agents are perfectly morally motivated. If they are not, then they might decide to exploit more vulnerable counterparties. My reply is that I am not assuming that agents are perfectly motivated. Instead, I am claiming that when specific circumstances obtain, agents that are motivated by their rational interests would have no reason to exploit others. This is the case for a firm facing a perfectly (or close to perfectly) competitive labour market, and for a political society conceived of as a well-ordered people. In addition, even if we relax the assumptions about background circumstances I have just made, but hold on to the idea that societies are sufficiently just and face favourable conditions, then exercises of power are unlikely to be successful insofar as those over whom such power would putatively be exercised would have sufficiently good outside options.

The second is that RW’s discussion plausibly refers to international labour markets. Assuming that even in ideal theory we would have immigration controls, then some workers could still be trapped in domestic labour markets that constrain their outside options. As a reply: I believe that what matters is if workers’ existing outside options are good enough, not whether their counterfactual ones could be even better. Good enough for what? Good enough for workers not to be subjected to exercises of power by their current employers. And I think that if each society is internally ordered by something like justice as fairness, and faces reasonably favourable conditions, this would obtain. That workers’ outside options could be made even better in a world where each society is fully just and labour is perfectly (internationally) mobile is not, in my view, a necessary condition to avoid discretionary exploitation.

4 The State and Trade Justice

A significant part of RW’s book is tasked with assessing the idea of trade justice from the perspective of states understood as agents within the global trade regime. In what follows I shall concentrate on Chapter 8, where they address the WTO. According to RW (2019, 139), states ‘ought to found an organization to pursue trade justice’. However, they also maintain that ‘while multilateralism is a requirement of justice, the WTO is exploitative’ (RW 2019, 139). Their thesis, in a nutshell, is that we do indeed need an institution that has the same function as the current version of the WTO, but that the latter is exploitative. Much of my discussion in this section of the essay will be a challenge to RW’s arguments in support of the conclusion that the WTO is exploitative.

Why is the WTO exploitative according to RW? There are two distinct complaints that they put forward. First, the WTO fails to perform well vis-à-vis other, non-trade related, moral concerns (RW 2019, 149). Second, the WTO is exploitative by reason of the way it governs the trade regime itself (RW 2019, 140). I shall leave the former complaint to one side (I largely agree with RW on that score) and concentrate on the latter, and especially on the idea that the WTO fails to fairly distribute the gains from trade.

What arguments support the conclusion that the WTO is exploitative because it allows for power-induced failures of reciprocity? According to RW:

compared to more forceful post-war proposals to integrate developing countries – think of the NIEO – the spirit of the WTO is to let developing countries have some share of globalization in return for playing by the rules that mostly maintain access for developed countries. The WTO has not benefitted the poor as much as it could have. Power has not been deployed to build a regime where genuine reciprocity reigns. (RW 2019, 140, emphasis added)

What is the evidence for this very broad and sweeping claim? RW’s major concerns are that, within the WTO system: (a) developing countries are losing money from developed countries’ protectionism and, relatedly, that agricultural exports should be less subject to trade restrictions; and (b) that protection of intellectual property rights has made reverse engineering of developed country technologies, and especially medical products, more difficult (RW 2019, 140).

I see several problems with RW’s account. To begin with, mentioning the fact that developing countries have not gained as much as they could have from the WTO (emphasis added in the quote above) seems puzzling. By that standard, the same would apply to developed countries, or to any agent joining any institution in any plausible set of circumstances.

Interpreted more charitably, the point must be that developing countries have gained less than they should have. But the latter claim can only be substantiated by providing a determination of how much a developing country should gain from membership in the WTO. And RW do not provide it. Put differently, if exploitation involves a power-induced failure of reciprocity, we should have a pretty clear sense of what reciprocity requires if we are to judge that a given institution is exploitative. Perhaps the thought is that all the gains from the trade made possible by the WTO should go to developing countries? Or that it would be enough if most of those gains went to developing countries? We simply are not told.

RW’s (2019, 151) argument could be rescued by referencing the idea that the WTO should have a development mandate. If we interpreted, as RW do, a development mandate as a mandate to enact policies that favour developing countries, then some progress could be made in assessing whether or not there are power-induced failures of reciprocity. For example, one could say that all arrangements that do not favour developing countries, and thus, that do not distribute the gains from trade unequally (i.e. more gains to developing countries) are ipso facto exploitative.

Even if we accept this kind of picture, one important issue remains unresolved. How can we know that there has been a violation of reciprocity if we lack a clear way (a metric) to understand what the net benefits of trade within the WTO system are and how they are distributed? To cite an UNCTAD report saying that developing countries were losing billions annually to protectionism soon after joining the WTO does not amount to particularly compelling evidence (see RW 2019, 140; United Nations 1999). First, it would be interesting to know compared to what baseline developing countries were losing those billions. I doubt the claim would stand if the benchmark is ‘no trade’ or ‘trade in the absence of the WTO’, but then we should ask why the benchmark should be ‘perfectly free trade’. Second, when protectionism by developed countries occurs, then both developing and developed countries lose from it and there is no a priori way to establish the relative magnitude of the losses. To illustrate the problem: if by protectionism we mean tariffs, then total deadweight losses aside, who pays how much of the tariff depends on the elasticity of demand for the good(s) in question, not on who is legally required to pay it (see Krugman et al. 2017).

However, the problem runs deeper than that. RW (2019, 52) say that they believe in comparative advantage. But the gains from trade extend well beyond comparative advantage (as determined by labour productivity or factor endowments). They include things such as economies of scale, increased input and product variety, and technology diffusion (Panagariya 2019). And these benefits are harder to measure, not to mention apportion between WTO members.

The last item on RW’s (2019, 140ff.) list of complaints about the WTO (technology diffusion) is worth singling out as it is explicitly discussed in the book. It is true that the protection of intellectual property rights favours companies in developed countries (see Hoekman and Kostecky 2001). Yet, the problem is not necessarily protection per se (which is widely recognized to be required as an incentive to invest in research to develop goods whose marginal cost of production is zero, but see Boldrin and Levine 2008). Rather, the problem may lie in the level of protection that is afforded to intellectual property. Excessive protection hurts all users of the protected products (raising prices unnecessarily) and all countries (stifling further innovation more than it encourages it), not just poor consumers and developing countries.

Furthermore, going back to my previous concern, even conceding that developing countries ought to receive special treatment when it comes to intellectual property created in developed ones, the lack of a precise benchmark simply does not allow us to draw any specific implication, unless one is prepared to argue that fairness to developing countries mandates no protection for intellectual property. A conclusion I would find implausible unless both ‘developing country’ and ‘intellectual property’ were defined much more restrictively than what RW seem to have in mind. After all, I see no obvious reason why the engine of a washing machine should be reverse engineered for free in, say, Kenya.

Finally, let’s consider the lack of market access for agricultural products from developing countries (that is, products where developing countries have historically enjoyed a strong comparative advantage) and its relationship to a putative WTO development mandate. Elsewhere in the book RW (2019, 160) extensively discuss the other main form of protectionism that affects the agricultural sector, namely, export subsidies to farmers in developed countries. In both instances, they correctly remind us of the fact that trade restrictions on agricultural products penalise producers in developing countries and consumers in developed ones.^[5] RW (2019, 161) also point out, once again correctly, that these restrictions favour developed countries’ exporters and foreign consumers. RW (2019, 161) then conclude that ‘consequentialist considerations support trade liberalization’, though they concede that the magnitude of the gains involved in trade liberalisation may vary significantly.

The more difficult question, however, is what to deduce from these facts when it comes to development. And this is because, as RW (2019, 12, 151f., 161f.) duly acknowledge, development is not determined by free trade alone and might even, at times, be in tension with it. After all, unless that was true, there would be no point in mentioning infant industry protection; the case for which RW (2019, 52) seem to largely accept.

We can be even more specific. The basic question is whether it is a good thing, in the long run, for developing countries to specialise in something that they currently enjoy a comparative advantage in. RW are clearly aware of this problem. For example, they quote an UNCTAD report suggesting that developing countries could be ‘locked’ in forms of comparative advantage that make efficient use of existing resources, but do not allow those countries to benefit from ‘dynamic productivity gains’ (RW 2019, 152). And by ‘dynamic productivity gains’ the report presumably refers to the development of comparative advantage in other sectors.

The main concern I have here is not with the overall picture that RW paint (I find it very reasonable). Whether there is a real case for infant industry protection, for example, is not something that has been unequivocally settled by professional economists and thus it would be nothing short of unfair and implausible to expect RW to do so. Rather, the point is that RW want to have it both ways. One cannot claim that the WTO is exploitative because, inter alia, it penalises developing countries in a sector where they have a comparative advantage and, at the same time, acknowledge that adopting a development strategy that relies on comparative advantage in that very same sector might be counterproductive. If the former is true, then the WTO is not doing developing countries any favours in allowing developed ones to artificially prop up their agricultural sector. If the latter is true, and if the goal is development, then making trade in agricultural products fairer would channel more resources to the wrong place (see Collier 2007). At the very least, RW should consider the possibility that in some cases making the WTO putatively less exploitative might weaken its ability to fulfil its development mandate.

Allow me to address one important objection.^[6] The objection is that I am failing to distinguish between two different kinds of claims: (a) that RW’s conclusions about the WTO are unsupported; (b) that the merit of their conclusions about the WTO are incorrect. This objection is particularly useful as it allows me to restate and clarify the overall argument I have advanced in this section of the paper.

My emphasis, in the main, has been on arguing that the claim RW make about the WTO (i.e. that it is exploitative) is unsupported. The basic problem I have tried to highlight is that they never seem to offer a clear benchmark for when a reciprocity requirement is violated between WTO members; I think that is a significant issue for an account that defines trade exploitation as a power-induced failure of reciprocity. However, I have also hinted at the fact that grounding an assessment of the WTO on the idea of a ‘failure of reciprocity’ is not the best way to expose the institution’s moral faults (and, to reiterate, this is perfectly compatible with accepting that the WTO has major moral flaws, see Garcia 2018). The reason being that failures of reciprocity within a complex institutional framework such as the WTO are difficult to establish.

Given the space available here, I cannot fully defend my wider thesis. I can only sketch a few considerations that seem to speak in its favour. Some of the concerns I have are as follows. To establish failures of reciprocity one would need: (a) a relatively precise account of the gains each member (or classes of members) of the WTO ought to receive from it; (b) a way to establish what counts as a gain (or a loss) resulting from WTO membership, and relatedly, the timeframe (e.g. short run vs. long run) in which such gains and losses are to be assessed; (c) a credible way to apportion the gains and losses made possible by the institution; (d) an account of how to trade-off different (often heterogeneous) gains and losses across the whole range of provisions hosted within the WTO framework. None the aforementioned tasks seems theoretically easy to carry out. Clearly enough, that something is not easy does not make it theoretically impossible or uninteresting (let alone wrong). Yet, it is worth noting that the challenges faced by RW’s approach seem significant.

5 The Corporation and Trade Justice

As RW (2019, 187) tell us, ‘states and companies are the most important actors in the domain of trade’. RW (2019, 188) theorize the existence of the firm, following Coase, as a way to minimize transaction costs, and explain the distinctiveness of the corporate firm based on the ideas of legal personality (the actions of the corporations cannot be reduced to those of workers, managers or investors) and limited liability (which protects the personal assets of those who have a financial stake in the firm) (RW 2019, 189). From a normative point of view, according to RW:

firms have trade-related obligations – most importantly their treatment of employees – and their relationship with outsiders, including communities where they conduct business, contractors and suppliers with whom they cooperate, as well as governments under whose jurisdiction they operate. Such obligations can be captured in terms of the requirement not to exploit. Companies ought not to use power to induce failures of reciprocity, but ought to transact on terms that satisfy transaction-specific claims of other parties. Companies also ought to see to it that transactions occur on non-exploitative terms. (RW 2019, 187)

The content of these obligations is further specified in relation to three principles: (a) to refrain from violating demands of justice; (b) to respect demands of justice; and (c) to support the fulfilment of the demands of justice that apply to the relevant agent (in this case, the firm) (RW 2019, 192f.).

RW discuss several important issues in connection to what firms ought to do or refrain from doing, both in general and within an integrated global economy, in order to avoid exploitation. In what follows I shall mainly comment on their account of the relationship between exploitation and working conditions broadly defined.

RW’s (2019, 193ff.) discussion of working conditions arise in two main ways in the book. Indirectly, when they discuss whether or not firms can be let ‘off the moral hook’. Directly, when they discuss the idea of a fair wage (RW 2019, 214ff.).

Let us start from the indirect approach. When discussing whether firms can be let off the moral hook, RW (2019, 193) offer a rebuttal of the idea that market pressures make it difficult for firms to respect demands of justice. Let us call the argument they are criticizing ‘the market pressure argument’ (MPA). So, how do RW reply to MPA? In essence they suggest that: (a) when production costs are a small part of the price, MPA is unconvincing; (b) productivity can be enhanced by non-exploitative working environments; (c) raising wages can be achieved by cutting costs elsewhere in the firm; (d) MPA does not pose significant mutual assurance problems because compliance costs with moral demands (such as safe working environments), even when unilaterally incurred, are small compared to the (moral) costs suffered by those who are exploited; and (e) even if MPA does apply we might be better off if some firms went out of business. As far as I can tell, the emphasis in all the aforementioned replies is on higher, non-exploitative, wages and/or better working conditions. Since RW have not yet specified what a fair wage is (something they tackle later on in the book), it seems fair to say that their remarks in this context purport to show that MPA does not justify exploitative working conditions broadly defined.

Are they correct? My concern is not so much with the conclusion they reach (namely, that MPA does not work), but with their arguments in support of that conclusion. More specifically, I think that the strongest version of MPA assumes something close to perfect competition and thus that the best reply to it must rely on the fact that perfectly competitive markets are hard to come by. Whether this kind repost to MPA is successful, all things considered, is a matter more complex to adjudicate (more on this below). After all, some would be inclined to say that, from a social point of view, if markets are not perfectly competitive we should not ask companies to offer better working conditions, but rather make markets more competitive. Yet challenging the assumption of perfectly competitive markets is, I believe, a necessary element of any plausible reply to MPA.

The closer the world gets to the perfectly competitive market model, the less the arguments RW offer against MPA hold water. To illustrate, in a perfectly competitive market firms sell their product at a price that is equivalent to marginal cost. If they did not, other firms would do so, and undercut them. Inefficient ‘non-cutting’ firms would simply go out of business (since they would literally be unable to sell any of their product). So, cutting costs ‘elsewhere’ does not seem like an option: if there were costs to be cut, they would be cut. The ‘other expenditures’ a firm incurs go to buy other factors of production or other versions of the same factor (e.g. more skilled labour). And, assuming perfect competition, the price of those factors is given and equals their marginal product, not some arbitrary amount that one can decrease at will. Thus, the idea that raising wages can be achieved by savings on other expenditures is dubious, no matter how small or large a proportion of production cost wages are.

Would it not be possible for firms in a sector to collectively reduce their profits to pay higher wages? If, once again, we assume perfect competition and we conceive of firms as rational actors, the answer is that it would be irrational. For, when perfect competition occurs, firms make so-called normal profits. Normal profits, or the absence of so-called economic profits, just means that firms receive a return on their investment that is equivalent to the opportunity cost of the capital employed by the firm (see for example Cowen and Tabarrok 2009/2018). Any less than that and firms would simply decide to change the kind of business they do, since, in other sectors of the economy, returns to capital invested are higher. So, unless we want to force firms to do what they currently do, the answer to the question asked at the beginning of the paragraph is ‘no’.^[7]

What about non-exploitative wages raising productivity levels? If this was the case, then in perfect competition, we would already have higher wages. Or, in a less compressed form, if there was good enough evidence that better working conditions always enhanced workers’ productivity to a sufficient degree, then firms would, ceteribus paribus, surely want to adopt those better working conditions.

Before moving on to the discussion of fair wages, let me clarify my argument about RW’s discussion of MPA.^[8] My sense is that we should distinguish three points. First, if markets are close to perfectly competitive, then MPA is close to being true. Second, in most circumstances, however, they are not. And this leaves open the possibility that MPA is not really a decisive argument against higher wages and/or better working conditions. Third, we should consider the following question: should we act in such a way that markets do become perfectly competitive? Here, my sense is that the answer is ‘it depends’. Most economists would say ‘yes’. But this is because they see perfect competition as a way to maximise social welfare in the aggregate. Even when they do take into account distribution, they often employ some version of the Kaldor–Hicks efficiency criterion without necessarily bothering to explain whether compensation is possible, or politically feasible (see Dietsch 2015; James 2012). In short, it seems plausible to say that, unless one adopts a welfare consequentialist perspective, there is no final (all-things-considered) reason to accept a requirement to make markets perfectly competitive. If that is the case, then my argument has the following logical structure: RW fail to consider the strongest version of the MPA; if they did, their conclusion (i.e. that firms should not be let off the moral hook) would still be valid, but the reasons they would need to offer to substantiate such conclusion would have to be different.

Let us now move on to the idea of a fair wage. RW start their discussion by both criticizing two specific ways of condemning wages as too low (mainly pointing out that the relationship between wages and prices is not a central one) (RW 2019, 204), and by arguing against the moral justification of market wages based on neoclassical price theory (RW 2019, 205f.). Both arguments, in my view, are pretty compelling. Things become more complex, and far less obvious, when we get to their own account of what a fair wage actually is. And clearly, that is a crucial topic for the book’s project. For, if trade-specific exploitation is a power-induced failure of reciprocity, then we need to say something about what reciprocity requires between firms and workers (wages being a central piece of the puzzle). One of my complaints about RW’s claim that the WTO is exploitative was precisely that we lacked a clear way to establish when demands of reciprocity were violated. While firms might fail to reciprocate workers in many ways (see RW, 223ff.), with an account of just wages on hand we would be much closer to solving the problem of what reciprocity actually requires.

How do RW proceed to establish their account of fair wages? They start by explaining the process of wage determination in standard neoclassical terms, and thus point out that, in equilibrium, there will be a producer surplus for workers, and a consumer surplus for employers (RW 2019, 207–208). They then suggest that whether wages are just or not will depend on how such surplus is divided, adding that the wage level determines that division (RW 2019, 208; see also Kates 2018). They then claim that failures of reciprocity might be a matter of cost incurred or contributions made (RW 2019, 208f.). Their argument does not amount to a specification of what precisely a just wage is, but it nevertheless purports to offer guidance on what determines relevant failures of reciprocity that would arise from unfair wages.

When it comes to costs, their claim is that ‘co-operators have claims against others in virtue of incurring costs of cooperation’ (RW 2019, 209), and that wages might be considered exploitative if they do not accurately reflect the costs that are incurred by workers. They then draw upon Reiff’s account of fair wages (2013) and argue that workers should receive the true social cost of their work (which includes not only subsistence needs but all of what Reiff calls contextual basic needs, including insurance for retirement and against incapacitation). They then claim that ‘where wages are insufficient to cover these costs owing to the employers’ power, there is a power-induced failure of reciprocity’ (RW 2019, 209, emphasis added).

Let me start by addressing RW’s analysis of the cost side of fair wages, for I believe that RW’s proposal leaves some important questions unanswered. For example, and in line with my comments above, it would be interesting to know what RW think should happen when market wages are below wages as determined by social cost and labour markets are fairly competitive. Standard economic theory tells us that unemployment would result (see Minimum Wage Study 2021). This would be particularly troubling precisely in those circumstances in which the argument for raising wages is likely to be more salient, namely in poorer countries; that is, countries where income from employment is crucial for workers’ standard of living since public welfare provisions tend to be absent.

However, as many discussions concerning the introduction of minimum wages have shown (see Card and Krueger 1994), unemployment might not result from wages higher than the going market wage if we assume employers to have some monopsony power. Further assuming that we can model the employer in those circumstances to operate as a non-discriminating monopsonist (i.e. it has to pay the same wage to all employees with the same level of skills), then the case for higher wages is clear since higher wages would be compatible with higher (rather than lower) employment.

Here, note that the setting of a higher wage would have the effect of maximizing employment only if the higher wage did not exceed the wage that would prevail in the relevant market if the monopsonist did not enjoy market power.^[9] Yet, there is no guarantee that a wage determined by reference to social cost would be equal to the one that maximises (or even increases) employment. If the former is sufficiently higher than the latter, then RW’s account would still imply higher unemployment than there needs to be, and perhaps more than there currently is. In that case, RW should provide some guidance on how to trade off less employment for some with higher wages for others. If, on the other hand, the wage determined by reference to social cost is lower than the wage maximising employment, then it would seem more plausible to recommend the higher of the two wages if what we are interested in is improving the welfare of poorly paid workers.

RW are clearly aware of problem I am highlighting and suggest that it can be taken care of through what they call the ‘price worth paying argument’ and the ‘stepping stone argument’ (RW 2019, 212ff.; see also Section 1 of this essay). The conclusion they reach is that neither justify exploitatively low wages. However, they imply that the trade-off should be handled by considering whether low wages credibly lead to development. They suggest that the link is shaky. I agree, but the main concern, I believe, is not the link between wage levels and development. It is, instead, the link between higher wages and higher unemployment in contexts where, as I have mentioned above, the state does not offer workers a robust social safety net.

So much so for ‘cost’, let’s move on to ‘contribution’. Partly taking cue from Dietsch’s work (2008), according to RW (2019, 209), ‘reciprocity requires division of a surplus in line with contribution, which we understand as a suitably modified version of time spent producing.’ And as RW correctly note, arguing for their thesis will require a defence of the moral significance of ‘time spent producing’ and a suitable account of how to measure it. Their justification for the moral significance of time spent producing is as follows:

Once we eliminate factors that are arbitrary from a moral standpoint within the context of cooperation, labor time expended is the factor most likely to survive. Unlike talent, productivity, or access to technology, time spent working generally is under individuals’ control and hence a factor for which, ordinarily, individuals are responsible. Within finite lives, time is arguable the most important resource. … For these reasons labor time should count as relevant contribution, generating claims of reciprocity. (RW 2019, 210)

In addition, RW want to suggest a ‘suitably modified version of time spent producing’, and thus the correct metric to evaluate the latter is not time sic et simpliciter, but time multiplied by the complexity of the task and the preparation and training that the carrying out the task require. In other words, ‘an hour worked may count as more than 1 h if time had to be spent acquiring skills’ (RW 2019, 210).

Here, I find a number of things to be unconvincing. To begin with, how long one is able to work might itself depend on factors that are arbitrary from a moral point of view. Some are able to concentrate for longer than others, some are physically fitter than others etc. Put differently, the assumption that work implies a trade-off with leisure for all workers in general tells us nothing about the specific nature of the trade-off for each distinct worker. And unless we assume all workers to have identical (mental and physical) assets to begin with, the trade-off is unlikely to be the same and might very well be determined by morally arbitrary factors.

Note that my criticism does not concern the proposed adjustment to time spent producing developing skills that allow workers to perform complex tasks. In the real world, opportunities to invest in human capital are unfairly distributed. Yet to ground one’s criticism of RW’s proposal on the latter observation would be unreasonable. For it would suggest that I am tasking their account of fair wages, by itself, to create a just society. Rather, my point is that if RW want to ‘cleanse’ wages from morally arbitrary factors, then time spent producing might be unable to achieve that goal, since time spent producing is itself partly determined by natural endowments that are distributed in a morally arbitrary way.^[10]

Second, note that RW’s proposal severs any kind of link between contribution and output. Rewards from cooperation are to be conceived of as a linear transformation of time spent producing (and social cost). But surely, time spent producing things no one wants ought not to be rewarded in the same way as time spent producing things that are in high demand. A link between time spent producing and wants, for lack of a better term, seems highly desirable from a social point of view, since it would send very clear information to those who want to spend time producing: you really ought to be producing X and not Y because most people want X as opposed to Y. The same applies, a fortiori, to the idea of attaching extra weight to the performance of tasks that require preparation and training. I see no reason why we should reward someone who spent a lot of time training and preparing for doing something no one is interested in.^[11] Relatedly, to equate contribution with time spent producing creates a less than promising set of incentives: the more time a worker takes to do something, the more, ceteribus paribus, they ought to be compensated.^[12]

Now, my objections might seem unfair. After all, workers on car factory floors and heart surgeons in hospitals, to take just two examples, do perform jobs that are useful, they usually do spend long hours working, and it is not implausible to think that the surgeon ought to be compensated more highly in view of the extra time spent on training and preparation. And this might be the picture that RW have in mind. Plausible as this picture may look, the problem is that it presupposes the existence of market prices and of the information (and incentives) that those convey. The fact that we are more likely to encounter car factory workers and heart surgeons instead of moon-singers and grass-counters is not a natural one. We can call this the knowledge problem (see Hayek 1945).

Two objections to the aforementioned criticism of RW’s contribution argument are worth addressing.^[13] The first is that it should be the job of managers to decide what to produce and that workers should not suffer if managers make mistakes on that front. I am unconvinced. To begin with, managers are workers too, and if they are to be rewarded according to the time they spend working at their task, they too should not be penalised if they make bad choices about what ought to be produced. Unless one wants to claim that the idea of a fair wage does not apply to managers.^[14] In addition, the point of raising what I have called, following Hayek, the knowledge problem, is not to assign specific responsibilities concerning who decides what to produce. In other words, when I say that strong distortions to the price mechanism for an important factor of production may have important informational costs, the point is not to say that workers or managers are responsible. Rather, it is to point out that, whoever we deem to be responsible for making those decisions, they would lack an important source of information; one that is connected to consumer preferences.

RW might legitimately ask why market prices are that important as a source of information. They might suspect that to think that they are one needs to believe that efficiency (understood as the maximization of a specific conception of social welfare) trumps all other considerations. I disagree. I do not think that my argument presupposes the acceptance of the idea that preference satisfaction has overriding value. What one needs to accept is that the allocation of productive resources according to consumer preferences has at least some value, and thus that an account of wages that seems to sever the link with those preferences is problematic (see Meade 1964/2013). Put differently, none of what I say implies that market prices should be the sole determinant of income distribution in a society (see Heath 2002; Meade 1964/2013). The problem is whether, or the extent to which, we think it is a good idea to affect the latter by changing the former (see Atkinson 2015; Thomas 2017).

6 Conclusions

RW’s account of trade exploitation as a power-induced failure of reciprocity, and their work more generally, is an important contribution to debates about fair trade. In this paper, I concentrated my critical efforts on the aspects of their work that I find less compelling. I have argued that their arguments to the effect that exploitation would still be present in ideal theory are unconvincing. I have also argued that their claim that the WTO is exploitative is unproven given that they do not offer a specific benchmark for what reciprocity requires between members of the institution. When it comes to exploitative working conditions, I have suggested that, even if their conclusion is largely correct, they ought to have considered a different version of the market pressure argument. On the topic of fair wages, I have emphasized that, on the cost side, their view would benefit from a discussion of the trade-offs between higher wages and unemployment in circumstances where income from employment is important. Finally, I have criticized the contribution side of their account of fair wages. The latter would require a more in-depth defence of the claim that time spent producing is not a morally arbitrary benchmark to evaluate contributions, and some appraisal of the absence of links between time spent producing and social welfare.