Authority and Agreement: The Case of Employment 

1. Employer’s Authority

An employer wields what I’ll call directive authority over their employees. Such authority has three aspects: (i) the boss’s power of command (ii) the boss’s entitlement to enforce their commands and (iii) the boss’s assumption of responsibility for compliance with their commands. We’ll consider (i) and (ii) in this section, and (iii) in the next. Let’s begin with the notion of command:

To command someone to do something is to communicate the intention of hereby putting them under a fresh obligation to do it, of obliging them to do it by the communication of this very intention (i.e. by declaration).

Legitimate directive authority involves the possession of a deontic power, a power exercised through command; where their orders bind this is because the speaker’s authority ensures the recipient is placed under an obligation to do what is commanded.

According to our definition, a commander purports to bind their subordinates by issuing a directive. Thus, they seek to get people to do what they tell them to do by deploying the idea of obligation. There are other ways to ensure that people reliably do what you tell them to do. The protection racketeer does not try to convince their ‘customers’ that they have any obligation to pay up; they simply make it clear that it would be wise of them to do so. Most employers do not operate like that; they issue commands in the expectation that their employees will feel obliged to obey. A boss enjoys de facto authority over their employees when the latter reliably do as they are told because they feel obliged to obey.¹ As we shall see, employers also have a battery of incentives and disincentives at their disposal to encourage their employees to comply but the smooth running of most firms presupposes that this is not the sole motive for compliance. The costs of constantly monitoring recalcitrant employees and enforcing instructions would often be prohibitive.

It is a further question whether the employer succeeds in creating a genuine obligation when they issue an order, whether their authority is de jure as well as de facto. Some authorities may be illegitimate and not a source of binding commands. Even legitimate authorities issue commands which do not bind for a variety of reasons, both procedural and substantive. I shall not defend the idea that your boss (or your government) enjoys precisely the authority they claim. What interests me is whether their commands sometimes change our obligations in the way they purport to, namely by declaration, rendering a subordinate obliged to do something simply because they have been ordered to do it.

We should be clear what is being claimed here, lest the point be too easily conceded. You might take someone’s command as a sign that they want you to do it and their wanting you to do it is as a good reason to do it. If they are wise their instruction may indicate that this would be a good thing to do on other grounds. Even if their instruction is not especially wise, should others go along with them it might be best if you do likewise. In each case, you can conform with the command for the relevant reason without accepting its legitimacy and perhaps much social conformity is of this sort. Whether or not our rulers care about why we conform with their commands, they do make the following claim: subjects without any extraneous reasons to conform have an obligation to do as we command which is furnished by the commands themselves, by the fact that we are communicating the intention to bind them.² I shall say that an employer possesses a power of command when this claim is at least sometimes true of their employees.

It is important to distinguish this power of command from the employer’s property rights.³ All owners enjoy proprietorial authority over their possessions. In particular, they can exclude others from the use of them or consent to their use only in specific ways. Now an employee very often needs to use resources owned by the employer in discharging their duties of employment e.g. when they move across the employer’s land or drive the employer’s van or use the employer’s computer. Here the employer’s proprietorial authority enables them to set the conditions under which these things are done. However, this is not the kind of authority which concerns us now for it does not enable the employer to positively direct the behaviour of the employee, to insist that they remain at work for a certain period and undertake specific tasks using the employer’s land, van and computer. That requires a contract of employment or, at the extreme, a claim to ownership of the person themselves. As a guest in your house, I accept that you can require me to leave at any point and may forbid me to enter a certain room but I do not expect to receive direction as if I were a household servant.

Your employer claims the authority to issue directives on a whole range of matters within the scope of your employment. They tell you where to work, what hours to work, what to wear at work, with whom you must work and what duties are associated with your role. Some of these things may be specified in your contract of employment but others are not and, in any case, most contracts of employment allow for them to be varied without the employee’s consent. The contract will also specify your place in the employment hierarchy, telling you from whom you must accept direction. Managers vary the requirements of your job by declaration, either on their own initiative or in obedience to higher authority.

Why is so much of our economy organised in this way? To function as a team, we must be able to coordinate effectively with our teammates, but does that require us to recognize a directive authority? Some orchestras manage perfectly well with no conductor. This works best when the ensemble is small, cooperative and/or lacks an especially skillful or knowledgeable member whom it makes sense to entrust with the direction of the whole enterprise. Firms are generally not like that. Employees have interests which conflict with the purpose of the enterprise and they also disagree about how the purpose of the enterprise should be pursued. To keep people in line, the boss must inculcate the idea that their employees are obliged to behave in ways that will promote successful coordination. Still, couldn’t this be achieved through agreements between members of the team acting as independent contractors? In our economy, such contractors do their job with no direction from the suppliers and customers with whom they trade. In principle, all production could be undertaken entirely through such market exchanges, in a series of one-off agreements whereby contractors buy inputs and produce an output that is sold to other contractors at the next stage of production. Why have firms with an internal hierarchy?

Ronald Coase observed that if production occurred solely through market exchanges, the parties would have to renegotiate the terms of the exchange whenever an alteration in market conditions occurred. The transaction costs of such renegotiation, including those involved in discovering what the relevant prices are, of negotiating, signing, and enforcing new contracts, and of any opportunistic behaviour by the relevant parties, are significantly reduced once we replace such exchanges with an administrative hierarchy in which a chain of managers wields open-ended authority to direct and redeploy workers as customers line up, machinery breaks down, co-workers call in sick, and other contingencies of production unfold.⁴ ‘If a workman moves from department Y to department X,’ Coase noted, ‘he does not go because of a change in relative prices, but because he is ordered to do so’ (Coase 1988: 35). Coase concludes that firms arise in a market economy just when the transaction costs of using market exchanges between independent contractors to regulate production can be avoided by replacing them with the intrafirm commands of an employer.

If Coase is right, hierarchy is needed for firms to discharge their social function; something similar will be true of public bureaucracies as also of many other organisations. Nevertheless, Coase allows that production is sometimes best arranged through agreements between independent contractors, as when I hire an electrician, a plumber, and a decorator to renovate my property. These people all agree to perform certain defined tasks for me within a certain timescale and for a certain price but how they do so is largely beyond my ken. They might ask me what I want and they need my permission to proceed but if the answer is not specified in the contract, they are free to refuse to do as I say without breaching their contractual obligations. Under certain conditions, they might feel bound to comply with my requests (the request is reasonable, important and so forth) but not simply because of our earlier agreement. That fact, and not merely the absence of transaction costs, is crucial to why many such people want to remain independent contractors.⁵ The benefits of organising economic activity around the firm notwithstanding, many wish to be ‘their own boss’ and not live under the direction of either a fellow tradesman or of their customers.⁶

Coase tells us that the difference between an independent contractor and an employee lies in the amount of discretion they are given over how to discharge their responsibilities (Coase 1988: 53-4). Yet a sub-contractor (for instance a factory farmer) may be hemmed in by a narrowly drawn contract specifying precisely what they must produce and how and when they must produce it. Whether this person would prefer to remain their own boss turns not solely on the density of the obligations they acquire but also on whether these obligations are the products of command rather than prior agreement.⁷ The self-employed are more comfortable with contracts of service than contracts of employment because whilst the employee is obliged to follow the subsequent directions of their manager, the contractor is not.⁸ Even a farm employee granted wide discretion over how they do their job enjoys this freedom only at the pleasure of their employer.

Not so many of us work on farms but most readers will have rented property. Consider tenancy agreements. They frequently go into great detail about the rights and obligations of landlord and tenant and this may be why rental contracts are (like contracts of employment) subject to a special legal regime.⁹ Nevertheless, they differ importantly from an employment relationship in that tenancy agreements as such give the landlord no discretionary power to issue instructions to their tenant. The tenant’s duties are standardly specified in advance in the agreement, need include no obligation to obey their landlord and so establish no directive authority.

This aversion to being ordered about also informs the behaviour of those who find themselves within a hierarchy of employment. My manager can instruct me to work with a certain rather difficult colleague but wisely refrains. Instead, they explain that I am the one person this colleague can work with, that their contribution is sorely needed and that other colleagues have their hands full. I can see that I ought to agree. Perhaps I also feel obliged to agree given that my manager has asked me but I may feel this way only because their request is so reasonable. I’m well aware that they could simply direct me to do it, a fact which informs my response, but it does so precisely because neither of us want to end up there. An order is a last resort.¹⁰

Where employment is at will on both sides, the idea that employees are more constrained than independent contractors may seem rather odd; isn’t it the other way around?¹¹ After all, if the employee doesn’t like an employer’s instruction, they can resign rather than obey it provided they are prepared to sacrifice their right to a salary, whilst someone who has contracted to paint my house must show up regardless of whether they are prepared to forego payment. True, employment need not be at will. For example, notice periods restrict people’s ability to resign their position, sportsmen are contracted to play for a team for a certain period, actors to perform in a certain film; these characters are obliged to remain at their post and accept direction. But even where one is free to leave, the preference for self-employment remains intelligible. An independent contractor is not expecting to have to obey orders in order to receive payment whilst someone who offers themselves for employment and who is being neither naïve, frivolous nor dishonest must expect exactly this. If an employee’s promise of obedience turns out to be empty because they resign as soon as they receive an instruction they disagree with, their employer will understandably feel misled.¹² By the same token, many self-employed people wish to earn their living without declaring their willingness to subordinate themselves to a boss.

How does the distinction just made between employees and independent contractors apply to the case of agents such as lawyers? A lawyer is obliged to carry out their client’s instructions, yet a ‘self-employed’ lawyer is in a different position from an ‘in house’ lawyer who is clearly an employee. First, the self-employed lawyer accepts their client’s instructions only within the scope of a specific and agreed objective—defending them against a particular charge for instance—whilst an in-house lawyer is obliged to take any case that their employer assigns them. Employers set for their employees objectives that are not specified in their contract of employment and do not confine their instructions to means. Second, in discharging their obligations, the self-employed lawyer uses their own resources (office, associates, equipment etc.) rather than those of their client and so their client is not entitled to control how they use these resources qua owner. By contrast in-house lawyers (and other employees) typically rely on the resources of their employer to discharge their responsibilities and so are subject to the proprietorial authority of their employer.

We have here a spectrum of subordination with employees at one end, independent contractors at the other and self-employed agents somewhere in between. These familiar facts notwithstanding, it remains a question why people have a special aversion to being ordered about. I’ll address this issue in the next section. In the meantime, let’s turn to the second aspect of directive authority: enforcement. Both our political rulers and our employers claim the right to enforce their directives, to ensure that what is commanded will happen even if their subordinate is inclined to do otherwise. We must distinguish two types of enforcement. One is the use of brute force as when a demonstrator is physically removed from a road or their property is confiscated to pay a fine. Here the result is secured by the application of force to body or property without the agency or cooperation of the coerced. With a few narrowly circumscribed exceptions (e.g. parents and teachers) the modern state claims a monopoly on the use of brute force; other social authorities who wish to invoke it (e.g. to enforce a contract) must apply to the state and its agents.

Though employers are (no longer) entitled to physically discipline their employees, they can coerce them in other ways. Disobedience is discouraged by the prospect of various sanctions—unpleasant assignments, unfavourable working hours, pay cuts, demotion and ultimately dismissal—or the denial of benefits such as holiday entitlement or promotion (González-Ricoy 2022: 11-12). If my builders don’t do what they agreed to do, I must go to court to enforce the contract (similarly with my self-employed lawyer) whilst my employer is entitled to back up their directives with a range of administrative measures. It is not just that they can modify my working conditions in ways I may dislike. Coercion involves my employer’s threatening to do so as a method of getting me to conform to their (other) commands. Here my aversion to the prospect of sanction is part of the plan—they are manipulating me—and I might object to my employer threatening to cut my pay if I don’t do as I am told, even if I wouldn’t object to a similar warning of pay cuts dictated by market conditions.

Those with the power of command generally do claim the right (together with some ability) to enforce their edicts, yet these are very different ways of getting someone to do something: whilst command places you under an obligation to do it, coercion either renders disobedience imprudent or bypasses your agency altogether. Furthermore, even if my employer can order me to wear a tie to work, that does not settle whether they are entitled to enforce their command. Such enforcement might be harmful to me or disruptive to the workplace and so they might be obliged to let the matter drop even when their command binds. How then are command and coercion connected?

Here is one hypothesis:

Enforcement: an authority is ceteris paribus entitled to force you to do something (should you be inclined to disobey) simply because they are entitled to command you to do it.¹³

Your employer can threaten to dock your pay if you don’t wear a tie to work because they are entitled to order you to wear a tie to work. Countervailing considerations may make enforcement unwise or even impermissible but there is a pro tanto entitlement to enforce binding commands. By contrast, one is never entitled to enforce a binding request simply because the request binds. Say I ask a friend to lend me their car. It may be that the considerations which render that particular request binding (its great urgency etc.) also establish that I am entitled to take the car should it be refused but, to put it mildly, that need not be so. Though a friend’s request may bind as a point of loyalty, the value of loyalty as such provides no basis for the idea that there is any pro tanto claim to enforce an obligation of friendship. No countervailing considerations are needed to explain the unenforceability of a request for help from a friend. We’ll suggest an explanation for the truth of Enforcement in the next section.

In this section we’ve examined the first two aspects of directive authority—command and coercion—and shown that employers claim directive authority over their employees. In the next section, I’ll argue that this means assuming responsibility for what they do in obedience to their orders. That connects the power of command both with the entitlement to coerce and with the third aspect of directive authority, the assumption of responsibility for compliance.

2. Employment and Responsibility

I’ll begin with a general conjecture: one can impose an obligation by declaration only if one also assumes responsibility for compliance. As we’ll see, the conjecture is confirmed by both promise (where the person bound by declaration is oneself) and by command (where the person bound by declaration is another). We must first explain what ‘assuming responsibility for compliance’ involves.

‘Responsibility’ comes in at least three forms: imputability, culpability and liability. Assuming liability means bearing responsibility for the consequences of compliance, for instance assuming an obligation to pay compensation for harm done. Employers are often held liable in court for the torts of their employees even when the employer is not personally at fault. Legal theorists disagree about the basis of such ‘strict liability’. According to one view, the employee’s tort is imputed to the employer (Stevens 2007). On another, there is no such imputation, rather liability depends upon the fact that employers are in a position to control the actions of their employees (Holmes 1891). Since I am not directly concerned with liability, I confine myself to noting that generally agreed facts about employer’s liability seem consistent with what I shall say about the forms of responsibility which do concern me.¹⁴

The ‘assumption of responsibility’ by an employer for the acts of their employee is first and foremost an assumption of imputability (Owens 2024). If an act is imputable to me then it is up to me to justify it, to ensure that it is what ought to be done; if an act is imputable to me and the act is wrong then I act wrongly in performing it. Many assume that where I have direct awareness of and control over the movements of a certain body, it follows that those movements are imputable to me as my actions. I don’t deny that these bodily movements are, in an important sense, my actions. However, it does not immediately follow that they are imputable to me, that it is up to me to justify them and that, if they can’t be justified, I am vulnerable to the relevant forms of criticism. Where my wrongful act is correctly imputed to someone else, I do not act wrongly. This is precisely how it is when I obey a binding but misguided order from my employer. Here the act is mine and furthermore I have the capacity to take responsibility for it. Nevertheless, it is up to my employer to justify it, to ensure that it is the right thing to do. I need not be acting wrongly even when I know that what I do is the wrong thing to do.¹⁵

Imputability is the basis for culpability. Someone is subject to criticism for an act where it is imputable to them and they cannot justify it. Such criticism takes various forms: an act can be wrong in virtue of being rude, silly, imprudent or irrational and can incur anger, disgust, contempt or first-person shame. Since we are focused on obligatory acts, the relevant forms of criticism are blame, resentment, indignation (or first-person guilt) and the relevant form of responsibility is culpability. When a wrongful act is imputable to my employer that explains why they are culpable and I shall use phrases like ‘assumption of responsibility’ or ‘passing the buck’ to refer to shifts of both imputability and culpability.

In claiming that obedience to a binding command involves an assumption of imputability by the authority, I am not just saying that culpability for the act commanded is shared between employer and employee. That may be how it is when my employer asks me to fire my colleague without ordering me to do so. Perhaps I feel bound by their request, perhaps I don’t, but either way I retain responsibility for my compliance, or rather I share that responsibility with my employer. And the employer may proceed in this way precisely in order to ensure that the decision which I am expected to execute is at least in part ‘up to me’. Once I am given an order, the situation changes: my employer assumes the entire responsibility for my compliance and complaints can be referred to them.¹⁶

Not every command of an employer binds their employee: if my boss tells me to drive the company van at twice the legal speed limit in order to make a delivery on time, I am not obliged to obey and should I decide to comply, my act is imputable to me. Nor am I obliged to obey their instructions about things altogether outside the scope of my employment e.g. my marital arrangements.¹⁷ In general, it remains the responsibility of the employee to determine exactly when a decision is a matter for their employer. Mistakes on this score are imputable to the employee but once this is settled, they can pass the buck to their employer. Discussions of authority and responsibility often focus on extremely foolish or profoundly immoral orders, orders whose bindingness is contentious. For my purposes, it suffices that some orders of some employers do bind us even when they are significantly wrong or misguided, as when I am told to dismiss someone who may not deserve to be fired. Here I retain responsibility for how I execute these orders—for tactfully conveying the bad news—but the fact that I am firing them at all is down to my employer.

Where the employee acts within the scope of their employment but without instructions or even contrary to the orders of their employer, there is no assumption of imputability by the employer for the actions of their employee. True, the employee is obliged to obey their employer and so the employer may be subject to criticism for their failure to exercise control over their employee, either by issuing a suitable instruction or by enforcing an instruction already issued. That failure will sometimes render the employer culpable for the misdeeds of their employee but that does not get the employee off the hook: culpability is shared not transferred. The employer may be liable for the actions of a disobedient employee even where they are not culpable (e.g. when their legal representative makes a contract against their wishes) but the employer need feel no guilt about what was done except in so far as they could have prevented it.¹⁸

It should now be clear why acting as an independent contractor is so different from acting as an employee. My contractors are obliged to fulfill the terms of their contract and may also feel bound to honour my reasonable requests but none of this involves their being obliged to obey my orders and so none of it involves an assumption of responsibility by me for their fulfilment of these obligations. I do share in the responsibility for the work of my builders but they cannot simply pass the buck to me for fulfilling the terms of their contract. Were they to become my employees, I could impose further obligations on them by declaration and, in accordance with our conjecture, that would mean assuming responsibility for their compliance.

Some may think that there can be no assumption of responsibility here because an employee freely undertakes the obligation to obey when they sign a contract of employment and therefore retains responsibility for any subsequent acts of obedience, even though the relevant obligations are imposed by the employer. This does not follow. One may be responsible for giving someone the opportunity to act badly without being responsible for their wrongful actions (as when one enters an area known to be dangerous and gets mugged). Where imputability for one’s own acts will be assumed by another, this is so even though the wrongs you anticipate are to be committed by yourself. The employee retains responsibility for having joined the firm and so for undertaking a general obligation of obedience but what they must answer for here is the decision to enable their superiors to assume responsibility for their own compliance with binding orders. That decision may have been justified even if some of the acts subsequently committed by them and imputed to their employer are wrong.

The idea that we are not responsible for actions performed in obedience to binding orders explains why a manager might feel relief once they are ordered to fire someone who needs to be let go. But doesn’t that feeling also throw doubt on the idea that we have a general interest in having our actions imputed to ourselves? Why should one wish to remain ‘one’s own boss’ if that means carrying the can for all that one does? True, adult responsibility can be a burden and occasionally we are content to lose it but even where avoiding the psychological cost of carrying the can is the greater gain, a self-respecting adult won’t value the opportunity to relinquish responsibility for its own sake. It remains somewhat shameful to be ordered about, to be treated as a person whose own view of what you ought to do should not determine what you do.¹⁹ And this applies not just to important decisions but also to one’s haircut; people resent being someone else’s charge like a child.²⁰

I’ve been outlining the connection between possession of the power of command and assumption of the responsibility for compliance. What about that other aspect of directive authority, the power of enforcement? Command and coercion are very different ways of persuading someone to do something. Nevertheless, employers often claim the right to enforce their commands because they claim that their employees are obliged to obey them. In the present section, I’ve been arguing that if your boss is entitled to give you orders, they must also be entitled to assume responsibility for your compliance. Now suppose that at least part of what is objectionable about coercion is that it deprives you of responsibility for doing the thing you are being forced to do. Then the fact that an employer is entitled to assume responsibility for the acts of their employee by issuing commands would indeed help to explain why they are also entitled to enforce their commands.

The idea that coercion deprives the coerced of responsibility for compliance is highly plausible where coercion takes the form of brute force, bypassing the subject’s agency altogether. But, as already noted, contemporary employers are not entitled to use brute force to make their employees comply with their instructions; rather they threaten various sanctions to forestall disobedience. Many writers regard such incentive threats as influencing what you do by providing you with a reason for compliance. If so, such threats do not compromise responsibility since you still get to decide what to do by making a judgement about what you ought to do. I suspect that this justificational model of coercion can’t fully account for the prima facie wrongfulness of coercion. I would argue for an excusatory model according to which someone who gives in to coercion can often deflect blame onto the coercer without being able to justify their compliance.²¹ If so, incentive threats do indeed affect whether we are expected to act on our own judgement about what we ought to do and so compromise our responsibility in a way analogous to command.

We now have a complete description of the directive authority that employers exercise over their employees. What forms of choice might establish the legitimacy of directive authority so understood? How does an employee undertake employment with all its ramifications?

3. Forms of Agreement

In the modern world, employment is voluntary for both parties, a choice embodied in a contract. In this section, I shall ask what form of agreement on an employee’s part might vindicate an employer’s demand for obedience.²² I shall consider four possibilities: (1) Choice, (2) Consent, (3) Promise, (4) Acceptance of Agency. These labels can be understood in various ways and my aim is not to capture ordinary (or indeed legal) usage. Instead, I give them each a specific sense with a view to marking some important distinctions.

We’ll begin with Choice. Often the fact that I control whether a social rule applies to me is a reason for deeming me bound by the rule should I fall within its scope. For example, I am obliged to dress formally at a funeral in part because I have chosen to be there. Here we are not concerned with exercises of ‘deontic power’, with choices directed at the normative situation, at people’s rights and obligations. Rather I’m using the word ‘choice’ to refer to decisions, intentions etc. aimed at the non-normative situation (e.g. a decision to enter a certain building), decisions with normative significance (Owens 2012: 168).

Let’s assume that the rules to which I have chosen to subject myself are not simply whimsical or oppressive, that they serve a genuine interest. Still, my having a real choice here might be necessary for them to bind me. Dressing formally at a funeral makes sense but I’m not bound to go to the trouble and expense of wearing a dark suit unless I have freely chosen to be there. Having such a ‘free’ choice involves several elements. First, I must know of the dress code or at least be aware that such a code may exist and be able to discover it. Second, I must be capable of avoiding the funeral without bearing exorbitant costs. Third, there must be no deception or coercion. Let’s assume these and other relevant conditions hold; then I may well be bound to dress formally having ‘tacitly agreed’ to do so.

In our funeral example, the rules in question are not the product of command. Let’s now suppose that I am confronted by a porter who requires me to put on a tie before entering a clubhouse because they judge that formal dress requires a tie. I may have good reason to comply with the porter’s demands even if I could somehow get into the club without doing so. For example, the porter might create a scene. Furthermore, given that I choose to try to enter the clubhouse, it might be the case that I am bound not to cause disruption or embarrassment by defying the porter’s instructions. But here it is not the porter’s command alone which binds me and my choice gives them no directive authority over me; rather the effect of my choice is to turn the extraneous reasons for doing as they say into an obligation. The need to keep the peace obliges me to comply given that I tried to enter.

A porter who claims directive authority over entrants to the clubhouse maintains that their issuing a command on this point obliges you to comply, regardless of whether the peace will be disturbed or anyone will be upset should you disobey. Here the porter takes on a special burden of justification, going beyond what would be needed to explain our subjection to sensible customs like a dress code. Compare the rules of friendship. Friendship is a positive feature of our lives, in part because it is a form of human involvement that we choose. We knowingly get into situations where bonds of loyalty will be formed and there are customs of friendship laying down what one should do for one’s friends, customs which bind us in part because friendship is voluntary. Yet, all of this notwithstanding, commands are foreign to friendship. Friends can ask for help and I am sometimes obliged to agree but a friend’s request for help never binds me in virtue of being the communication of an intention to impose an obligation.²³

Might an exercise of a complementary deontic power like Consent be required to ground the power of command? As I shall understand it, to consent to someone’s A-ing is not to wish or intend or choose that they A, rather it is to communicate the intention of hereby permitting them to A (Owens 2012: 165-72). Thus, unlike choice and like command, consent is directed at the normative situation and operates by declaration. When I permit you to cross my land I communicate the intention of hereby making it the case that you are no longer obliged to refrain. Consent must do more than that if it is to be a source of directive authority: it must impose an obligation on the consentor. Perhaps what is really needed is for me to promise to (continue to) permit you to cross my land, to grant you an easement. A promise imposes an obligation on the promisor and it does so by way of a declaration directed at the normative situation. Furthermore, there is a clear sense in which any binding promise puts the promisee in authority over the promisor. Once the promise is made, the promisee has the power of release (i.e. the power to unbind the promisor by declaration) whilst the promisor has no power to release themselves from their promise.²⁴

The classical political philosophers were contractarians, maintaining that political authority originated in a (collective) promise but the argument of the last section suggests that the agreement which establishes a directive authority must be a promise of a special kind. An ordinary promise gives the promisee control over what you owe them without subordinating you to their commands. As we saw, my builder may be happy to promise me various things and they stand ready to fulfill their promises without being prepared to become my employee and obey my instructions. To explain what more is involved in promising to obey someone, I shall introduce our final form of agreement: the acceptance of agency.

We take our cue from employment law which treats the employee as the employer’s agent.²⁵ Recall our conjecture: one can impose an obligation by declaration only if one also assumes responsibility for compliance. The agent-principal relation has two complementary aspects. First, the agent is obliged to obey the instructions issued by the principal within the scope of their agency. Second, when they obey, the principal ‘owns’ their actions and the agent can pass the buck for compliance to them. That is true of employed agents such as household servants and in-house lawyers; it is also true of non-employee agents such as lawyers who accept the instructions of their clients. Neither feature applies to those who simply make a promise e.g. to drive someone to an airport. Here the promisor both retains the right to decide how to accomplish this objective and retains the responsibility for fulfilling the promise. They may share that responsibility with the promisee but they cannot pass the buck.

I propose that the special promise of obedience which employees make to employers involves an acceptance of agency. As such it is the counterpart of the exercise of a higher-order power first noted by Hobbes, one which he calls authorization. For Hobbes to authorize X to A is to permit X to make use of your power or your right to A.²⁶ The principal authorizes the agent to make use of their power to admit people to their home on their behalf or to sign a contract in their name on condition that their instructions are obeyed, all in accordance with our conjecture. In accepting this agency, the agent is promising to facilitate the principal’s use of their rights and powers e.g. by being present to let people into the house or being available to negotiate and conclude contracts on their behalf.²⁷ This promise restricts the agent’s physical activity in the same way that promising to drive someone to an airport would but because it is an acceptance of agency, it also transfers responsibility for the activities involved in fulfilling the promise to the principal.

The power to accept agency is no more mysterious than the power of authorization and is indeed required for the latter to have effect. On the other hand, if what was said in the last section about the need to ‘be one’s own boss’ is accurate, it is a question what value is served by the exercise of either power. People are often willing to take on the role of agent in return for compensation and are responsible for agreeing to do so but it may be doubted whether this option should be made available to them. In the final section, I’ll address this issue by comparing employment with two other relations of subordination: (voluntary) servitude and state authority.

4. Employees, Subjects and Slaves

A master claimed to own their slaves; a modern employer does not claim to own their employees. What is the difference here? Both relationships may originate in an agreement (formerly people sold themselves into slavery or agreed to enslavement to avoid capital punishment). Doubtless no one would enslave themselves unless the alternatives were dreadful but the fact that a promise was made to extract you from a terrible situation does not by itself invalidate your promise. Both employment and slavery involve granting the superior the power of command. True, the authority claimed by the employee is much more restricted than that claimed by the master over their slave (Locke 1988: 322-3). Furthermore, unlike the modern employer, masters felt entitled to use brute force to enforce their edict and yet these differences look like matters of degree.

To reveal a difference of kind here, we need to dig below the normative surface and uncover the rationale behind the powers claimed by the master and those claimed by our employers, the interests they are meant to serve. Slavery is a form of ownership and I would argue that the rationale distinctive of ownership is to facilitate people’s domination of their environment, enabling them to identify with and control their chattels as they do the parts of their own body (Owens 2022: Chapter 7). Since the slave is a person, dominating them involves the master’s turning the slave into their deontic satellite: the master stands in for the slave in the nexus of rights and responsibilities, depriving them of moral personality, of any independent normative standing.²⁸ Let’s suppose that’s correct. Aren’t firms also owned by individuals and aren’t its employees parts of the firm? Should we conclude, as some anarchists have done, that employment is a milder form of servitude?²⁹

We can avoid this outcome by maintaining that even if the employer owns the firm, it does not follow that the employer owns the employees. Ownership of the firm is ownership of a thing, a legal entity. The owner is entitled to dominate that thing by controlling the rights and obligations possessed by others in respect of that thing e.g. by making contracts in the name of the firm either with other firms or with individual suppliers and customers. Among those contracts will be contracts of employment and this is how an internal hierarchy is established. The owner either sets themselves up as the manager or appoints a manager and that manager is then authorized to issue instructions to the other employees. If the manager is distinct from the owner, then the owner can issue binding instructions to the manager as an employee and the same is true of the instructions which the manager issues to other employees. At no point is authority within the firm based on the idea that one individual owns another.³⁰

What then is the basis for the authority of an employer? Earlier we explained why a hierarchical firm is a useful social form by reference to its ability to marshal information efficiently and facilitate the process of production. This hypothesis left it open whose interests are meant to be served by such efficiency. If the employer owns the firm, that may entail that the firm should be run in the interests of the employer. On the other hand, wider social interests may also be relevant here: the interests of the employees, of other stakeholders or even the welfare of the economy as a whole (Owens 2022: 190-6). However that may be, what undergirds employer’s authority is the fact that a hierarchical firm serves some interest other than the interest in inter-personal dominance which underlies slavery.

Our hypothesis places limits on the scope of the employer’s authority and on the reasons for which they can exercise it (Kolodny 2023: 39-44). The employer can’t issue instructions about matters that have nothing to do with the smooth running of the firm. Such instructions are usually ultra vires and do not bind the employee. In respect of matters that do fall within their jurisdiction, the employer is obliged to consider only the functioning of the firm when formulating their orders. For example, they can fire someone because they are superfluous to requirements but not because they don’t like their religion. Still, as we shall see, the employer might succeed in firing an employee (i.e. their dismissal might be valid) even though they wrong them in dismissing them. Similarly, orders issued on inadequate or irrelevant grounds may wrong the employees precisely because they also bind them.

How is the exercise of employer’s authority meant to ensure the flourishing of the firm? I said that the employer comes to own actions performed by their employees in the course of their duties of employment in that they assume responsibility for those acts. However, the rationale behind this assumption of responsibility is quite different from that underlying the idea that the slave acts in the name of their master. Authority within the firm is meant to secure efficient operation of the firm and the firm couldn’t function effectively if each employee followed their own judgement of what was best for the firm. That is why unjustified dismissals and foolish orders are not invalidated simply because they are bad. Disagreements about the merits of a line of action are dealt with by transferring decision-making from the employee to the manager.³¹ By the same token, the employee who must obey an order of which they disapprove relinquishes responsibility.

Consider a firm owned by its employees. Here the employees (qua shareholders) can appoint (and dismiss) its managers, say at the Annual General Meeting. They can also lay down the general direction of policy which the management are to pursue. On the other hand, between AGMs and qua employees, they are obliged to accept the direction of the managers they have appointed in respect of how they should contribute to the implementation of that policy. Supposing that the point of those directives is to further the interests of the firm’s owners, if the managers are incompetent or worse, they wrong the employees qua owners when they issue faulty instructions. That doesn’t absolve the employees of their obligation to obey–there would likely be chaos if individual employees could ignore the instructions of the management whenever their own judgement differed–but it does mean that simple disobedience does not personally wrong the manager (qua manager).³² If anyone, it wrongs their fellow employees/shareholders.

Turning now to political authority, mightn’t the rationale for employer’s authority differ too much from that of the authority of the state for the comparison to be helpful? True, both employment and government are meant to ensure the flourishing of an artificial entity–the firm in the one case and the state in the other–but, at least in the modern world, there is a huge difference. A manager who issues orders in a firm is either the owner of the firm or the firm’s agent. The modern state is owned by no one and its officials are meant to pursue the good of the citizenry and not any narrower set of interests. Given this, why think that agreement will play the role in legitimating the state that it does in legitimating employment? It would be no surprise if consent were necessary for the latter but not for the former.

I introduced the example of a firm owned by its employees with a view to modelling the state as imagined by the classical political philosophers. Just as people may be both shareholders and employees, so the same person can be both citizen and subject. In a contractualist state, the citizens institute a sovereign who is authorized to resolve disagreements about how their collective life should be run. The sovereign and their officials are obliged to issue orders which further the interests of the citizenry and they wrong the citizens should they do otherwise. Mechanisms may be put in place for the dismissal of wayward or incompetent officials, yet this cannot be a daily occurrence; so long as they are in office, the citizens, qua subjects, are obliged to do as they say (Gauthier 1969: 175). The sovereign should declare war only if this serves the interest of the citizens but, qua subjects, the citizens are obliged to go to war should the sovereign command it.

How to think about the allocation of responsibility here? We’ve suggested that firms acquire employees when the latter agree to become the employer’s agent. Many people may join the firm at once but they both arrive and depart as individuals. That remains true when the employees are also the shareholders. As we saw in the last section, employees remain responsible for joining the firm (as for leaving) but they can pass the buck to their employer for individual acts undertaken in obedience to their binding instructions. On the contractarian model, the same is true of the state. There would be little point in having a sovereign unless they had a large measure of discretion in determining, say, when and against whom to declare war. When I go to war in obedience to the sovereign’s command, I can pass the buck to the sovereign for what I do though, in the contractarian state, I remain responsible for putting myself in a situation where I am obliged to obey their orders. Thus, we can model authority and responsibility within the contractarian state on that within a firm owned by its employees; there could be a state whose normative infrastructure worked in this way. Whether any actual state fits the contractarian model is another matter. Where we are born to political obligation and cannot walk away with any ease, the model won’t fit.

Can the agency model of employment we have outlined at least vindicate the authority of the boss? People become employees by making contracts of employment, contracts which are frequently voluntary. Though most of us must work for our living, in a modern economy we can usually choose whom to work for. Furthermore, the option of self-employment is often available, an option with no clear analog in the political realm. Even those who doubt that employment is genuinely free under capitalism should agree that, in ideal conditions, the acceptance of agency may be voluntary and so able to legitimise employer’s authority. Anarchists might continue to insist that contracts of employment, however voluntary, are no more binding than contracts of servitude but, given the differences with slavery already noted, the burden of proof lies with them.

To sum up, many of us suppose that an employer can have legitimate authority over their employees provided those employees have agreed to be employed. In this paper, I’ve analysed both the kind of authority that a contemporary firm claims over its employees and suggested a form of agreement which might legitimate that authority. The exercise of an employer’s authority over their employees involves an assumption of responsibility by the employer for the acts of their employees, explaining why many people are reluctant to seek employment even though they stand ready to enter into other contractual relationships. I maintained that the form of agreement needed to ground such obligations of employment is the acceptance of agency. Finally, we considered the rationale underlying the authority of employers, distinguishing being employed both from being owned and from being subject to the authority of a government. Much more might be said but we have done enough to explain how, in the modern world, employment is a contractual relationship.³³