Domination at Work: Resisting the Radical Diagnosis

Abstract Political theorists and philosophers have converged mainly on the following claim: workers face the threat of domination in the workplace. While many propose reforms to capitalism to address this domination, a growing chorus of radical Republicans argue that domination at work results from the capitalist distribution of ownership. In this paper, I argue that this focus on ownership is a mistake. Instead, I contend we ought to direct our attention to the role that managerial discretion plays in the firm's structure. My intention in advancing these two claims is to carve out space for a position that acknowledges the domination of workers, but does so in a way that is compatible with private ownership in the means of production.


Introduction
Most of us either work or have worked in jobs where our livelihoods depended on the choice of the boss. If they decide one way, we will be able to get on with our lives. However, if they choose the other way, we are not entirely sure how we will carry on. In this regard, most of us have experienced what it is like to be at the mercy of another in the workplace.
For political theorists and philosophers, this fact will remind them of the concept of domination. Domination, like many political concepts, is used in a variety of ways. 1 A particularly influential account associated with republican tradition ties domination to arbitrary power (Lovett 2010;Pettit 1997). More specifically, person A is subject to domination when they are subject to the discretionary power of another person, B, in such a way that B's discretion (i) does not need to take into account A's practical reasoning and (ii) implicates A's basic interests (see also McCammon 2015). This account suggests a prima facie case for the domination of workers.
Republicans of all stripes agree that domination at work is a problem. Disagreement mostly surrounds solutions. While one candidate may use workplace rules and regulations to check arbitrary power (Dagger 2006, 162-3), another shores up workers' power of exit as a form of antipower (Taylor 2017, 46-65). Yet another possibility gives workers formal control over how their firm is run (Anderson 2017, 65-70;González-Ricoy 2014, 244-8). Here, think of co-determination in Germany, where workers have representatives on boards of directors. While these solutions require reform to the status quo, they remain compatible with a broadly capitalist system. For this reason, however, many find these solutions inadequate.
A growing group of theorists I call radical republicans agree that workers are dominated. However, unlike more moderate republicans, these radical republicans argue that the domination of workers is built into the capitalist system (Arnold 2017;Bryan 2023;Cicerchia 2022;Gourevitch 2013;Gourevitch 2015;Muldoon 2022;O'Shea 2020a). If capitalism is the cause of domination, the solution requires rejecting private ownership of the means of production. Ultimately, by linking worker domination to capitalism, these radicals often revise republican understandings of domination, making the role of structure explicit. Because of this, we address domination at work by changing the structure: we give workers ownership over the means of production. Or so these radical republicans suggest.
The radical republican claim has gone mostly unchallenged. A few raise concerns about radical republicanism, suggesting that republican concern with domination does not fit well with radical politics (Kandiyali 2022). Others take issue with radical republicanism, arguing that republicans need not align with radical ends (for example, Frye 2020; Taylor 2019). This paper seeks to contribute to this latter line of criticism by challenging the radical republican diagnosis of domination at work. I believe radical republicans focus far too much on the distribution of ownership as a source of domination. Against this sourcing of worker domination, I advance two arguments. First, I claim that many arguments that link the unequal distribution of ownership to the domination of workers face problems that have yet to be addressed in the literature. Many radical republicans do not appear to fully appreciate that owners are a diverse set of actors with interests that do not neatly align, providing opportunities for checking their power within the system of private property. Second, I point out that a significant source of worker domination is orthogonal to the distribution of ownership in the economy: managerial discretion within the hierarchical firm. Many have previously noted the relationship between managerial discretion and worker domination (Anderson 2017;González-Ricoy 2014, 238;Hsieh 2005, Hsieh 2008. However, my claim is that this threat of domination is independent of the question of firm ownership, which has not been sufficiently emphasized in the literature. Some even invite linking these two issues, such as Hsieh, who sees the solution to managerial prerogative as the system of 'property-owning democracy'an ambitious economic system which uses heavy state intervention to disperse ownership over productive assets widely (Hsieh 2005, 129-30). 2 Others, such as Anderson, focus so much on defending their right flank they remain vulnerable on the left (for example, O'Shea 2020a, 557-9).
My intention in advancing these arguments is not just to raise difficult questions for radical republicans, but also to carve out space for a position that acknowledges the domination of workers without being committed to seizing the means of production. Because of this, I do not think my arguments against the radical republican position are conclusive. Nonetheless, they need answering.

The Radical Republican Claim
Radical republicans argue that the domination of workers is a basic feature of capitalism (Gourevitch 2013, 598-603;O'Shea 2019, 16-19). Under capitalism the means of production are privately held. Importantly, these assets are distributed so that not everyone owns productive assets. Thus, a capitalist system is roughly defined by the relationship between two groups of people. There are those who have ownership over the means of production, capitalists, and those who do not, workers.
The problem with this arrangement from the radical republican perspective is that workers only own their labour power (or their capacity to work). If workers only own their labour power, they must find a way to meet their basic needs. Lo and behold, capitalists are willing to pay workers for their labour power. Because of this, workers must work for capitalists to meet their basic needs (see O'Shea 2019, 18-19).
To be clear, workers are not forced to work for any particular capitalist. The labour market allows workers to leave their job and seek employment elsewhere. Capitalism is not the same as feudalism (Halliday and Thrasher 2020, 9-14). However, workers must work for 2 On the idea of property-owning democracy, see the collection of essays in O'Neill and Williamson 2012 and Thomas 2017. a capitalist. Taken together, capitalists control the terms by which workers earn the wages they need for subsistence and whatever goods they may need to flourish. Because of this asymmetry between the position of capitalists and workers, the radical republican argues that capitalists dominate workers. As Gourevitch puts it, 'domination appears when this structure is so unequal that some group of owners privately controls all of society's productive assets. Those who do not own are economically dependent on employers for jobs, wages, and thus their livelihoods ' (2013, 602). Summarized, this is the basic radical republican argument in its broadest shape: (1) Workers lack access to the means of production.
(2) Capitalists control access to the means of production.
(3) Workers must access the means of production to acquire wages to satisfy basic needs.
There are many ways of filling out the argument. Some are more compelling than others. Nonetheless, radical republicanism involves two commitments in my telling: (i) unequal distribution of ownership of the means of production leads to the domination of workers, and (ii) eliminating this domination requires restructuring the economy to address this distribution. The conjunction of these commitments constitutes the radical republican claim, or at least the version of the claim I am interested in.
There are at least three ways of filling out the relevant notion of domination in the radical republican argument: agential, class-based, and structural. Briefly: Agential Domination: Capitalists dominate workers in that a particular worker at a firm is subject to the arbitrary power of the particular capitalist(s) who own(s) that firm.
Class-Based Domination: Capitalists dominate workers in that workers as a class are subject to the arbitrary power of capitalists as a class.
Structural Domination: Capitalists dominate workers in that capitalism as a structure empowers capitalists over workers.
These are not mutually exclusive accounts of capitalist domination. You can find strains of each in radical republican authors who draw on multiple forms of worker domination (for example, Gourevitch 2013, 598-603;O'Shea 2019). Importantly, these forms of domination are said to often act in support of each other. For example, one might argue that the structure of capitalism puts the capitalist class in the position it is in, and this position grants individual capitalists arbitrary power over individual workers. First, however, we need to separate each claim. If we can work our way from agent to structure, this will force us to be crystal clear on where exactly the arbitrary power lies in the capitalist system. This is my focus over the next three sections.

Agential Domination
Some worker needs money to put food on the table. This worker goes out and finds a job at a factory owned by a capitalist. Once at work, however, the worker finds himself pushed around by the owner, realizing that he will be fired if he does not do as the owner says, so the worker does what the owner says. This looks like the story of how a capitalist qua capitalist dominates a worker qua worker. The worker only does what the capitalist says because he fears termination. The capitalist only has this power over the worker because she owns the factory where that worker works.
I do not doubt that individual capitalists control access to certain means of productionthis is just what it is to be a capitalist. But it is not clear that controlling access to the factory, in our case, on its own, is sufficient for domination. Domination does not exist whenever an agent holds power over another person. Domination exists whenever an agent holds arbitrary power over another person. Arbitrary does not mean random; it is linked to uncontrolled, discretionary power. In our factory example, the owner holds power over the worker, but this does not entail that the owner holds arbitrary power over the worker. More needs to be said before we can conclude that. This is true even if our worker must work for some capitalist (though not necessarily the factory owner) to meet his basic needs. Suppose someone is forced by circumstance into a particular type of relationship, though not forced into any particular token of that type of relationship. In this case, it does not follow that that person is dominated. To establish domination, we need further argumentation linking being forced into a type of relationship (though not any particular token!) to arbitrary power.
Consider the following analogy: suppose you are suffering from a terminal but treatable illness. You must receive treatment from a doctor, or else you will die. In this situation, you are forced to see a doctor in the same sense that a worker is forced to work for a capitalist (neither you nor the worker has an acceptable alternative). Further, you are not forced to see any particular doctor, much like the worker is not forced to work for any particular capitalist. It seems extremely odd to say that you (and patients like you) are dominated by doctors by virtue of the structure of your situation. To be clear, the patient-doctor relationship is beset by an asymmetry of power and risks domination. However, to say a relationship risks domination is not equivalent to saying that the relationship entails domination. Further, to say a relationship risks domination does not entail eliminating that relationship. No one who worries about domination at the hands of doctors thinks we should get rid of doctors. Instead, they focus on ways to control the power of doctors (professional licensing boards, informed consent, etc.). This situation is different from the master-slave relationship, which is a paradigmatic case of domination. To be a master means having dominating power over a slave. You cannot 'check' the power of masters sufficiently without eliminating the whole class of masters. This is not true of doctors. Because of this, we need an argument to explain why capitalists are more like masters than doctors.
One difference between capitalists and masters is that workers are not bound to any particular capitalistthey have a formal choice in who they work for. Because of this, a capitalist cannot impose whatever conditions they see fit because those workers can consider other options (similarly, if you think your doctor is not advancing your best interest, you have the option of seeking out another doctor, assuming you live in a country with an adequate healthcare system). To be clear, a formal right of exit is insufficient to check capitalists' power (Malleson 2013, 613-21). However, that is not my claim here. My point is that, even if workers are forced to work for a capitalist, there are ways of leveraging other actors (including other capitalists) against individual capitalists, constituting an admittedly imperfect check on that capitalist's power (Lovett 2010, 53;Pettit 2006, 142;Taylor 2013, 595-7, Taylor 2017. Beyond exit, our worker could potentially appeal to labour regulations, form a union with other workers, threaten to quit, leverage an offer from a competing manufacturer, or fall back on the welfare system, etc., all as a way of constraining the power of the owner of the factory. I will not spend much more time on this agential story, as this is all well-explored territory. Radical republicans offer many responses to these ways of constraining capitalist power. However, in my view, the most powerful responses rely on the claim that constraining individual capitalists in these ways fails to take seriously how class-based and structural worker domination is. Thus, I turn to these less scrutinized lines of argument.

Class-Based Domination
The factory owner is a capitalist, but she is not alone. She is a part of a class of capitalists. When discussing the dominating power of capitalists over workers, perhaps it is a mistake to focus on the individual worker or the individual capitalist. Instead, one must zoom out and think about the relationship between classes. While the factory owner does not control the means of production all by herself, capitalists own and control the means of production as a group. Using Nicholas Vrousalis's phrase, capitalists dominate workers because they exercise power as 'a coalition of agents' (2021, 90-1; see also Vrousalis 2013, 139). Freedom does not check the power of capitalists; one faces capitalists wherever one goes. As Sam Arnold puts it, 'When capitalists speak in one voice, workers must obey' (2017, unimportant 120). In this development of the radical republican argument, the owner of the car factory dominates the worker in a roundabout way. It is not as an individual capitalist that she dominates the worker. Instead, the car factory owner taken together with her class dominates the worker.
Class is notoriously complicated, and I cannot address these complexities here. But let me seize on an ambiguity in the radical republican claim about what it is for capitalists to act as a class to create problems for the narrative that capitalists dominate as a class.
What does it mean to speak in 'one voice'? Collections of persons can dominate; that is not in contention. However, not just any collection of persons can dominate. Otherwise, non-domination would be impossible as we often find ourselves around people (Dowding 2011;Simpson 2017). Because of this, theorists of domination say that only group agents or teams, actual or capable, dominate (Lovett and Pettit 2018, 368-78). Group agents exist where some sort of organizational scaffolding produces common judgements, decisions, actions, etc. among a particular collection of persons (Lovett and Pettit 2018, 368; see also List and Pettit 2011). If speaking in 'one voice' means that capitalists act as a group agent, the claim is easy to dismiss because there is no such organizational structure that unites the behaviour of capitalists. More plausible is the idea that teams can dominate. A team does not require the same organizational structure as a group agent. Instead, a team simply requires shared awareness of a shared intention and the capacity to act on that shared intention (Lovett 2022, 32-3;Lovett and Pettit 2018, 377-8).
If capitalists act or are capable of acting together as a team to advance their interests at the expense of workers, I take this to be a form of domination. But we are not describing capitalists as capitalists in these sorts of cases but, instead, as colluding capitalists (see also Pettit 1997, 141). When capitalists collude, they create an illegitimate cartel that bypasses competitive market checks on their power. Importantly, collusion is not necessary to be a capitalist; it is but a possibilityindeed, anti-trust laws are built into our legal structuring of the market to combat this possibility. These do not just stop active collusion; they also deter companies from coordinating their behaviour in the first place. In this way, anti-trust laws help keep possible teams from being capable teams, mitigating domination.
Consider what happens when anti-trust law fails in this task. In the 2000s, California tech companies colluded against the interests of their employees. Companies such as Apple and Google shared an understanding that they would refrain from hiring one another's employees. This cooperation limited employees' ability to leave their firms, providing a clear case of employers exercising arbitrary power over their workers. However, the source of the domination in this situation is not the mere fact that their employers were capitalists but rather that these businesses colluded in anti-competitive ways to limit an important tool of workers to keep in check the power held over them: the labour market. Tech firms not involved in this collusion did not contribute to the domination of the workers at the hands of the illegal coalition. Lillian Cicerchia observes that a market system undermines collective action by promoting competition (2022,(11)(12)(13)(14)(15)(16)(17)(18). While Cicerchia sees this as a way markets enable domination, it is worth noting how much this is a feature rather than a bug (Heath 2014, 93-115). As the case of tech companies colluding against workers suggests, one group's collective action is another group's domination. Undermining the possibility of coordination is a valuable way of checking against domination.
But there is a different way to understand the claim that capitalists act as a class to dominate workers, one that requires no actual or potential explicit collusion. When Sam Arnold says capitalists speak in 'one voice', he does not say capitalists act as a cartel. Rather, '[a]cross a wide range of issues', capitalists largely agree with one another (Arnold 2017, 120). Specifically: Let labor market conditions be as competitive as you please, capitalists will still require workers to show up for work. They will still require workers to toil at a profit-maximizing pace, under profit-maximizing conditions, for a profit-maximizing wage. They will still force workers to lay down their tools at the end of the day and hand over the surplus created by their labors. And they will still reserve for themselves the right to decide what to do with this surplus (Arnold 2017, 120).
Capitalists share interests such that they will act as though they were colluding to enforce an episodic will upon workers. Under such conditions, competition will fail to check individual capitalist power because other capitalists offer the same terms, not out of collusion but instead out of a convergence of interests.
This is similar to a worry occasionally raised regarding price transparency among hospitals. The case for price transparency is that it creates competitive pressures that, hopefully, keep prices down. However, some speculate that price transparency may have the opposite effect in particular circumstances. If prices are public, this may allow hospitals to converge on higher prices without explicit collusion. This is not a mere theoretical worry. In the mid-2000s, residents of Martha's Vineyard, an island in Massachusetts, realized that gas prices were significantly higher on the island than compared to the mainland. These residents sued for illegal price fixing. However, the court found against them because there was no evidence of an agreement among these gas stations. The gas stations were engaged in tacit collusion: without any actual agreement, these four gas stations were able to coordinate on an oligopoly. The anti-competitive behaviour of the gas stations was made possible in part because there are only four gas stations on Martha's Vineyard, entry into the island's market was difficult, and prices were public. These features made cooperation on relatively high prices easy, as members of the oligopoly could police for defection and react in kind even in the absence of an agreement. Unfortunately for us, this coordination ends up having a dominating effect in the case of the tacitly coordinating gas stations. Importantly, such coordination suggests that even if the gas stations did not tacitly collude, they were capable of such collusion. Thus, domination exists in this scenario whether or not the gas stations end up cooperating. We might say something similar about the tacit coordination of capitalists against workers; whether or not they tacitly collude with another, they share enough interests to be capable of tacitly colluding against workers and thus dominate workers.
This argument is stronger than any argument that posits active collusion with capitalists as a whole class but suffers a similar flaw. It overstates the similarity of capitalists' interests and understates the severity of capitalists' collective action problems. Because of this, worker interests can align with particular capitalists against other capitalists in ways that undermine the capacity of capitalists to act as a team. Take one of Arnold's examples: imposing a particular wage on workers. In Arnold's telling, capitalists seek to push for an as low as possible wage, and this shared interest tells poorly for the workers. In reality, capitalists can and do offer higher wages than other capitalists to attract other workers, which seems to be in tension with the 'one voice' story. For example, Amazon has created a de facto $16 minimum wage in many areas in the US, much to the dismay of competing employers (Herrera 2021). Or, consider the expansion of more flexible working arrangements in response to a tight labour market during the pandemic (Fox 2022). In these cases, employers are not acting as 'benevolent masters', simply exercising their power in a way that benefits workers out of the goodness of their hearts. Instead, market conditions compel such employers to offer better terms than other employers. It isn't easy to make sense of these companies catering to worker interests at the expense of one another in Arnold's telling. Yes, capitalists share the interest of seeking 'profit', but we need more to justify the claim that this shared interest turns into coordinated behaviour that makes one class a team capable of exercising power over workers. The global economy is not the same as the market for gas on Martha's Vineyard. The conditions that made collective action accessible to the gas stations in the latter case do not hold for capitalists in the former: the global economy is not small, entry into the market is not limited, and the terms capitalists offer to workers are often not public information.
I have been focusing on the direct relationship between capitalists and workers, either through the firm or through class relations. But this might be a mistake. Explicit in Arnold's argument is that the capitalist system drives capitalists to act in particular ways (2017,12). This suggests that the source of domination is not so much class interests but, rather, something about capitalism that places workers in a position such that capitalists dominate them. This leads to the idea of structural domination.

Structural Domination and Capitalists
The claim that capitalism involves structural domination is woolly in the way that all structural explanations tend to bethere are various ways of understanding the relevant notions of 'structure' and how structures dominate. To make my task more manageable, I will only take on two accounts of structural domination in capitalism, the influential one of Alex Gourevitch and the more recent development of Gourevitch's ideas by Alexander Bryan. 3 Gourevitch argues that the rules of capitalism render workers subject to arbitrary power. In his words, 'It is the absence of reasonable alternatives to selling one's labour that guarantees propertyless workers have to sell their labour to some employer or another. That absence of alternatives is created by the distribution of control over productive assets, a distribution that is secured by legal protection of private ownership.' (Gourevitch 2013, 602). As I understand it, the claim is that capitalists dominate workers by supporting a distribution of property in the means of production such that some will control access while others will not. What differentiates a capitalist from a doctor is that the doctor has no prior relationship with what makes the patient need the doctor's services in the first place. By contrast, the capitalist is linked to a system that compels workers to work for some capitalist. To use Gourevitch's words, 'The laws protecting the current distribution of property arbitrarily removed the one reasonable alternative to working for another person. That, in itself, was a form of domination, exercised by the entire class of property owners and the state agents protecting them. ' (2015, 108). On this line, capitalists dominate workers, not by owning private property in the means of production but, rather, by maintaining a system where others lack property. Domination enters the equation because capitalists could always choose otherwise: capitalists could endorse a system in which they lack their privileges. The fact that capitalists could choose a different ownership structure is a form of arbitrary power. Or so says Gourevitch.
What does it mean to say that capitalists could choose otherwise? Is it the individual capitalist who can choose otherwise with regard to her particular piece of property, or is it that the capitalist class can choose otherwise? Either way generates problems for the radical republican claim. Suppose I own a factory; how exactly do I maintain the structure of private property in a way that dominates those particular workers in my factory? Is it in my retaining ownership rights over the car factory? Suppose the radical republican's answer to this question is yes. In 3 Both Bryan and Gourevitch keep central arbitrary power in their accounts of structural domination. I set aside accounts of structural domination that step away from the focus on arbitrary power (e.g., Roberts 2017, 93;Thompson 2018;Vrousalis 2013). To respect these arguments would require a different treatment that spells out the relevant notions of domination and how they differ from the neo-republican ideaa task I do not have space for here. that case, the problem is that if I cede ownership of the factory to my workers, the structure of private property remains intact. This is true even if my particular ownership claim has changed. But if my actions did nothing to impact the structure of private property, this implies that by retaining my ownership, I could not have dominated my workers. It would seem peculiar that I could dominate them by maintaining a structure of private property when no action I could take makes a difference to that structure. Whether or not private property in the means of production exists is not up to me. Given that discretion is a necessary condition for domination under the account we are considering, my lack of discretion is problematic for the argument that I dominate my workers. Perhaps the idea is that it is not me by myself that maintains the structure of private property. Rather, I, along with all the other capitalists, preserve the structure of private property. This seems closer to the truth. There is a structure of private property that requires maintenance to reproduce itself. However, maintenance as a term is ambiguous in ways that raise problems for Gourevitch's argument.
On the one hand, maintenance is something a person or group of people perform as an intentional action. For example, I maintain my car by following the recommended service guidelines. I do not think that is what is going on in the capitalist case. On the other hand, maintenance can refer to whatever set of actions keeps the system in place, regardless of whether that set of actions is under the discretionary control of a person or group of persons. Consider a traffic jam. A traffic jam has a structure to it, even if it is undesirable. Further, such a structure is maintained by the drivers' actions that make up the traffic jam. Nonetheless, no driver nor the group of drivers has anything we could meaningfully call discretion over whether the traffic jam continues (if only!). This seems closer to the relevant sense of maintenance under capitalism. However, suppose this is the relevant way capitalists maintain the distribution of private property. In that case, it is difficult to say that capitalists dominate workers because, just as the drivers in the traffic jam lack discretion, so do the capitalists. No discretion, no domination.
The traffic jam analogy might provoke the following response. It might be correct that no individual or subset has discretion over the traffic jam but consider an alternative case. What if some of the drivers benefitted from the traffic jam and actively wanted to continue such a traffic jam? Such drivers would have no interest in resolving the traffic jam. Similarly, capitalists who benefit from the structure of private property have no interest in changing the structure to one where they cede their privileges. Indeed, capitalists may want such a structure in place, even if capitalists lack the discretion to change that structure.
The problem with this objection is that a group of persons benefitting from a particular structure or even wanting that structure in place does not entail those people dominating others in that structure. 4 Again, we need an explanation of how those involved have the power to dictate that structure through their actions. If people lack the power to change the structure according to their discretion, then it is not clear there is domination at play as opposed to some other type of concern, such as oppression. 5 There is a difference between a structure placing one group of people under the discretionary power of another group and a structure that systematically advantages one group at the expense of another. An explanation of why capitalism is more like the former than the latter will have to respond to my earlier observations surrounding the difficulty of describing capitalists as a capable team. For these reasons, I am not convinced by Gourevitch's 4 In the revision process, it came to my attention that Frank Lovett makes a similar observation in his recent book, distinguishing 'condition takers' and 'condition makers' (Lovett 2022, 46-7). It is worth looking at Lovett's discussion as it is subtly different from mine. 5 It is telling that Gourevitch has appeared to abandon the idea that capitalism is a system of domination and instead opts to describe capitalism as a system of 'class-based oppression' in more recent work (2018, 907-9; 2020, 107-8). Perhaps concerns such as mine might explain this shift from domination to oppression. For an account of oppression amenable to this sort of analysis, see Ann Cudd's work (Cudd 2006, 23-6), but notice that Cudd herself does not think capitalism is a system of oppression (2006, 122-33). particular account of structural domination under capitalism. But there is another way to go structural that does not require positing such a power over structure to capitalists.
In a recent paper, Alexander Bryan updates Gourevitch's argument in a way that does not rely on the claim that capitalists dominate through their arbitrary maintenance of the private property system. Instead, Bryan argues that certain economic structures generate dependence. In Bryan's words, '[D]omination is best characterized not in terms of exposure to the possibility of particular kinds of interference but as a condition of dependence on the cooperation or permission of other agents to engage in productive activity ' (2023, 693). Non-domination requires meeting what Bryan calls the 'independent production criterion': when some agent, A, requires cooperation with others to engage in productive activity, these others will both (i) provide such cooperation and (ii) do so on terms that are controlled in the relevant sense by A (Bryan 2023, 699). To illustrate, in an economy where investment decisions by a company impact the livelihoods of workers in a particular town, and this company does not need to consider those worker's interests in its decisions, Bryan would say these workers are dependent on the company and thus dominated (assuming that the workers are unable to engage in productive activity through some alternative arrangement that reflects adequate control). Here, Bryan's work advances Dorothea Gädeke's recent arguments that domination is structurally constituted (Gädeke 2020a;Gädeke 2020b). An economic structure may not only disempower workers by limiting who has access to capital but also may place those disempowered workers under the discretionary power of the company in our simple case (Gädeke 2020b, 34-5). Unlike Gourevitch's argument, Bryan's argument has the advantage of not positing anything like discretion over the structure of private property to capitalists. Nonetheless, while providing a more plausible account of structural domination in the economy, Bryan's argument loses the radical bite of Gourevitch's argument.
Bryan's independent production criterion does not refer to the idea that workers must have an ownership stake in the means of production. All that it requires is suitable control over potential cooperators. Indeed, Bryan himself states, 'Other agents owning and controlling goods that others need to access to engage in production is not of itself a source of structural domination: that domination emerges in the absence of adequate constraints on those agents' use of those goods' (2023, 700, emphasis added). In this quotation, Bryan concedes that private ownership of the means of production does not lead to domination all on its own. It is only uncontrolled private ownership that generates domination. Control, as I have noted and Bryan himself notes, takes a wide range of forms, many of which do not involve a radical restructuring of the economy. To be clear, Bryan argues that natural monopolies or goods subject to network effects may require more radical solutions, such as direct democratic control and ownership, but this is a far cry from the claim that the solution to worker domination requires eliminating private ownership of the means of production (2023, 701). Moderate republicans will readily acknowledge that certain goods and services should be subject to democratic control and ownership. There may be dispute over which goods and services qualify, but such disputes will not be over whether we must abolish the system of private property. 6 They assume it. Thus, while Bryan gives us a clearer vision of an economy free of structural domination with his independent production criterion, such an economy need not be free of private ownership in the means of production. Because of these observations, I am sceptical that going structural has as much force as the radical republican thinks it has. Further, I worry that a focus on ownership distracts from a more apparent source of domination.
Before moving on to my preferred diagnosis of the domination of workers, let me make two caveats. The first thing to note is that I do not deny that capitalists can dominate workers. Indeed, I have accepted that particular capitalists colluding can dominate particular workers, and this is something we should combat. Further, one might acknowledge the role of political mobilization 6 There will also be disagreement over the effectiveness of democratic control in certain cases (for an overview of social democratic experiments in public management, see Heath 2019). through the state as a form of capitalist domination (for example, Klein 2017, 859;Preiss 2021, 493-501). My point in this section has been to argue that the idea of structural domination, at least when such domination references the concept of arbitrary power, is insufficient to establish the radical republican claim. The connection between capitalism and the domination of workers is not inevitable, unlike the connection between slavery and the domination of enslaved people. The second thing to note is that my arguments in this section do not necessarily vindicate capitalism from the perspective of non-domination. It may be the case that the best way of mitigating the domination of workers involves rejecting capitalism. This would be because some alternative economic form, such as socialism, would better protect workers against arbitrary power. I do not think this is true, but I want to be clear about the limits of my argument.

Managerial Discretion and the Domination of Workers
This is a good place to step back and look at the cases that motivate worry about the plight of workers. Consider the extended description of unfreedom at work by Chris Bertram et al.: On pain of being fired, workers in most parts of the United States can be commanded to pee or forbidden to pee. They can be watched on camera by their boss while they pee. They can be forbidden to wear what they want, say what they want (and at what decibel), and associate with whom they want … In addition to abridging freedoms on the job, employers abridge their employees' freedoms off the job. Employers invade employees' privacy, demanding that they hand over passwords to their Facebook accounts, and fire them for resisting such invasions. Employers secretly film their employees at home. Workers are fired for supporting the wrong political candidates ('work for John Kerry or work for me'), failing to donate to employer-approved candidates, challenging government officials, writing critiques of religion on their personal blogs (IBM instructs employees to 'show proper consideration … for topics that may be considered objectionable or inflammatorysuch as politics and religion'), carrying on extramarital affairs, participating in group sex at home, cross-dressing, and more (Bertram, Gourevitch, and Robin 2012).
What is striking to me about these cases is that the abuse of power is not by capitalists or, to use more contemporary terms, stockholders. Instead, as Bertram et al. explicitly acknowledge, the bosses or managers are responsible. However, the Marxist language of capitalists and workers tends to obscure the actual structure of the contemporary workplace on several dimensions. When the radical republican says that workers must work for some employer, the idea of an employer is nebulous. Who is the employer? The individual stockholder who has no knowledge of, or will ever know, the worker? Or is it the mid-level manager who decides to hire this worker and determines how that worker's labour will be deployed? This lumping together of managers and owners is a mistake. Bosses and managers are not the same as the owners of the firm.
In this section, I argue that the radical republican focus on the distribution of ownership is orthogonal to what I take to be the basic threat of domination to workers: the hierarchical nature of the firm in conjunction with a system of at-will employment. As my charge of orthogonality suggests, this threat of domination exists independently of whatever group owns the firm. Thus, the radical republican focus on the question of ownership over the means of production fails to fully recognize the more fundamental challenge of managerial discretion: how do we control the power of managers? Of course, ownership and control are not unrelated issues. Ownership is one way of establishing formal control over something (though formal control does not always turn into effective controlthis is particularly true for any individual worker). But these issues are importantly distinct, as I hope to show.
The radical republican may think the charge of orthogonality is itself orthogonal. Nothing stops a radical republican from simply responding, 'Yes, managerial discretion is a problem. But, given suitable responses to your earlier discussion, this does not give us a reason to think that ownership is not also a problem.' I agree that such a response is available. However, recall that my purpose in this paper is to resist the radical diagnosis of worker domination. To do so successfully, it is not enough to suggest problems for the radical position. I should also provide an alternative diagnosis that does not rest the responsibility for worker domination at the feet of private ownership in the means of production. Otherwise, I worry that I am telling the patient bleeding on the table that they are perfectly healthy. More critically, I want to explain why I think the radical story tends to occlude my diagnosis in practice, even if there is no logical contradiction between accepting the ownership story and accepting the managerial story. Even if the discussion so far does not ultimately convince the radical republican, I will be happy if I generate more conversation about how questions of corporate governance are of central importance to understanding domination in the workplace. 7 To understand managerial control, it is important to understand what the firm is. An important development in political theory and philosophy is seeing the firm as a social form distinct from the market (Anderson 2015;Ciepley 2013;González-Ricoy 2022;Singer 2018;Singer 2019). Following the economist Ronald Coase, the primary difference between the market and the firm is that the former coordinates economic activity through the price mechanism, whereas the latter coordinates economic activity through hierarchical relations of authoritythis is where managerial control enters the equation (Coase 1937;Singer 2019, 52-72).
Markets coordinate behaviour through the price mechanism, meaning that individuals contract with one another by assigning prices to particular goods and services. Assuming a suitably competitive market, no individual market actor has the power to influence this price, as other actors can always undercut attempts to do so. Everyone is a price taker, not a price maker. This feature of competitive markets makes them appealing to some republicans (for example, Taylor 2013Taylor , 2017. However, the price mechanism is not perfect, and we do not live with perfectly competitive markets. Coase famously pointed out that using the price mechanism to coordinate economic activity is not freetransaction costs abound (Coase 1937, 390-2). For example, negotiating a contract requires acquiring information about prevailing prices, taking time to bargain with the other party, dealing with uncertainty about the future, etc. As a result, Coase argued, there are times when it makes sense to opt for a different way to organize economic activity. This is where the firm comes in.
Unlike the price mechanism, the firm does not coordinate economic activity through the decentralized decision-making of individual agents. Instead, the firm is a hierarchical organization where some have the power to give orders to others. In particular, managers have openended authority to direct their subordinates as they see fit. This managerial prerogative bypasses the transaction costs of using the price mechanism. If I need you to complete this report today, I do not need to negotiate a new price, figure out if this is a good price in light of the competition, and so on -I simply tell you to get it done. Thus, an important feature of the firm is that managers have wide discretion over how they direct their subordinates. This discretion over subordinates raises the possibility of domination.
Managerial discretion over subordinates alone does not entail domination. However, managerial discretion does entail domination against a background of at-will employment. Under at-will employment, managers can fire their employees without a need to justify their decision. Because of this, managers do not simply issue orders. They also issue threats. Do this, or else I will fire you. To state the obvious, firing can significantly set back a person's basic interests under present 7 This may generate a different set of issues for the radical republican. I have argued elsewhere that once we appreciate the challenges of managerial control, it becomes less clear that radical changes to ownership alleviate managerial domination instead of exacerbating it (Frye 2020; cf. O'Shea 2020b).
circumstances. This power to fire in conjunction with managerial discretion looks like a case of dominationworkers who fear termination know that their fates are in the hands of their bosses. This is where republican theorists of the economy ought to focus their attention.
A clear benefit of thinking through the link between worker domination and managerial prerogative is that it reveals that the focus on firm ownership is, in part, a red herring. 'Ownership' of the firm is sometimes controversial; some suggest that all firms are unowned (Ciepley 2013, 145-7;Strudler 2016, 112-21). 8 In this context, ownership acts as shorthand for marking some group that (i) exercises formal control over the firm through an oversight board and (ii) is a residual claimant, or the group entitled to what money remains after all contractual obligations are fulfilled (Hansmann 1996, 12-16;Singer 2019, 9-10). In a traditional joint stock company, this means that formal control occurs through a board of directors elected by the stockholders and that stockholders are entitled to their share of the profits in the form of dividends. To show that a focus on ownership sometimes leads us astray, consider a non-profit such as a university.
No one owns such non-profits. Instead, the non-profit is formally controlled at the highest level by a board such as a Board of Regents, and all revenues exceeding cost are reinvested back into the non-profit (Hansmann 1996, 227-45;Heath 2010, 187). Now think of the position of the custodial staff at a university (let us set aside the case of tenured faculty, which presents an unusual case in the labour force for several reasons); staff are directed by managers within a university to act in a particular way. Unless some arrangement protects that staff member, they find themselves in a position of doing what the boss says or else. The whole radical republican story about how workers must work for capitalists entirely misses this. And yet, on its face, the situation of the janitor at the non-profit university seems structurally identical to the janitor at the privately-owned car factory. Both find themselves in the position that they do as the manager says or face severe consequences. There is the threat of domination, not in who owns the firm.
One might worry that a similar objection to that raised in my previous discussion about capitalist domination may raise its head here. Specifically, managerial discretion is not the problem; unchecked managerial discretion is the problem (Taylor 2017, 58-9). We often subject ourselves to discretion in ways that do not imply domination. To use Robert Taylor's example, ordering omakase at a sushi restaurant places my dinner choice in the hands of the chef, but no one thinks this discretion implies domination (2017,58). Similarly, we might argue that managerial discretion on its own doesn't create domination. Instead, managerial discretion, plus a background of policies, can give that discretion bite over a worker's basic interests (lack of adequate exit options, insufficient safety net, no representation in managerial oversight, etc.). Should that background change in the right ways, then managerial discretion loses its dominating character.
I agree that managerial discretion could, in principle, be checked in such a way as to limit or even eliminate managerial domination. That being said, my point is not that managerial discretion is a problem wherever it exists. Instead, I am trying to shift the conversation from one about private ownership to one about corporate governance. If there are moderately complex firms, there will be managers with prerogative, whether or not those firms are owned by capitalists, workers, consumers, etc. But, as long as there is managerial prerogative, workers face the possibility of domination so long as that discretion is insufficiently checked or implicates a worker's basic interests.
Notice that there is an asymmetry between the radical republican story about capitalists and mine about managers. Managerial prerogative only applies to workers once they begin working for a firm. The domination does not occur through the labour market. 9 If I am looking for a job 8 If these thinkers are right, what does this mean for the radical republican view? I am not sure. 9 I do not deny that domination can occur through the labour market, particularly under conditions that approach monopsony. My point is that the threat of domination by managers remains even under very worker-friendly labour market conditions. and you have a job to offer, you are not my boss. Instead, you are a potential boss along with others. If I do not like your offer, I can say no and continue my job search. Once I am in the job, the costs of that calculation change. You now have control over my career in a way you did not before. Given potential background policies, this may not matter so much, but under current conditions, it seems clear that this power is unchecked in ways that suggest domination. In contrast to the radical republican view, my view is that the threat of domination does not arise from the fact that people must participate in the labour market. Instead, it arises from the simple fact that managers have power over their employees.
One might wonder whether ownership may sneak in through the back door as the real source of domination. Specifically, if managers work for capitalists, then capitalists can dominate workers through managers. The capitalist has power over the manager, who then has power over other workers. Through transitivity, the capitalist has power over the worker.
This line of thought is mistaken for the reasons I will explain shortly. But, for now, let me concede the point: capitalists can dominate workers through the mechanism of managerial discretion. Even if I grant this, my claim in this section remains. Specifically, a very salient threat to domination rests in managerial discretion. Even in a world without capitalists, a moderately complex economy will have firms with managers who boss workers around. The radical republican story that focuses on the distribution of ownership over the means of production clouds this fact. This is true even if capitalists can wield this mechanism of managerial domination. But I do not need to concede the point.
The claim that capitalists dominate workers via managerial discretion involves the assumption that power is transitive between persons. If person A has power over person B, and person B has power over person C, then person A has power over person C. But this is not necessarily true (see also Lovett 2010, 69-71, 97). For example, consider a simple model of bureaucracy within a well-functioning democracy. Elected politicians have power over bureaucrats through their power to appoint to offices and legislate the limits of an agency. Likewise, bureaucrats have power over citizens through the powers of a given agency. But notice that citizens have power over politicians through elections. Whether this simple model accurately reflects how electoral systems interact with the administrative state is not important. What is important is that it shows that power is not necessarily transitive. Thus, if capitalists have power over managers and managers have power over workers, it does not follow that capitalists have power over workers.
Perhaps, more seriously, a tacit assumption behind the objection is that there is a tight link between the formal control granted by owners of a firm and effective control. But there is no such tight link between formal and effective controlfamiliarity with the history of business ethics scandals up to the present day suggest how often managers fleece stockholders, the purported owners and controllers of the firm. Attempts to tighten this link via financial incentives have historically underdelivered and even backfired. 10 There remains a gap between whoever owns the firm and the wide discretion granted to managers. This gap is important because managers have their interests separate from that of the firm as a whole and those they command. This gap presents opportunities for managers to abuse this discretion for personal gain. Consider the problem of sexual harassment in the workplace. Many women fear confronting sexual harassment from a manager because they are worried about the consequences. Sexual harassment is not just a problem for the women who experience it; it also ends up damaging the firm. Unsurprisingly, women who experience harassment are more discontent and less productive on average in their jobs (for example, Willness, Steel, and Lee 2007, 146-9). This is a stark example of how managerial discretion can act as a vehicle for bosses to abuse their power regardless of what formal control is granted to owners. Whatever power firm owners have over workers 10 The case of Enron is instructive here (Healy and Palepu 2003). More broadly, on the failure to reign in managers, see Bebchuk and Fried 2004. through managers is extremely vague and imprecise. The managers, however, have a very clear power over workers.

Conclusion
Radical republicans argue that workers are dominated because our current economic system involves a distribution of private property where some have control over the means of production and others do not. I have argued that this claim, in its most powerful forms, requires that capitalists operate as a team or a potential team; it is unclear whether the coordination required for such a team exists, nor does it appear readily at hand. More seriously, this approach to domination in the workplace obscures the more obvious threat: managerial discretion in conjunction with background conditions that give that discretion bite, such as at-will employment. Whether or not firms are privately owned, worker-owned, publicly owned, or owned by no one, managers will retain authority over subordinate workers. If we wish to avoid domination in the workplace, we must face this fact head on. Given that a major function of the firm as an economic form is to give managers discretion, it is difficult to check managerial prerogative and not introduce different kinds of problems. Nevertheless, this is the problem that proponents of freedom as a non-domination face.