One of Donald Trump's most (in)famous promises during the campaign that saw him elected as US President in 2016 was that he would build a wall along the US–Mexico border. Further, he insisted that Mexico would pay for this wall.

At the time of writing, it remains to be seen whether the wall will be built at all, let alone who will actually pay for it if it is. However, this is not my concern here. Rather, I am concerned with the legitimacy of this pledge. If we take Trump at his word, then he was proposing a policy that would place a significant tax burden on a group of ‘outsiders’ – Mexicans, neither citizens of nor resident in the United States. This goes beyond the authority that states usually claim to tax their citizens and/or income or wealth in their territory. Moreover, it violates the principle ‘no taxation without representation’ – a principle which led the American colonies to fight for independence from the British Empire.

Historically, the chief importance of parliament had been to give consent to taxes. This is why the franchise had been tied to property requirements, so only taxpayers were represented. The American revolutionaries considered it unacceptable that they were required to pay taxes to a government in London but were not represented in the British parliament. There were, of course, two ways that the revolutionary demand could be satisfied. First, they could have been granted representation, in which case ongoing taxation would no longer have been without representation. Second, if they were no longer taxed, then there was no need (at least as far as this slogan goes) to represent them in parliament.

In recent years, political theorists have become increasingly interested in questions about the composition of the demos (people) that rules in a demo-cracy. We think it undemocratic to exclude certain groups, such as women or ethnic minorities, from the franchise. Yet reasons given for including them – such as, so that they can protect their own interests – also look like reasons to extend suffrage still further than we do at present. In particular, in a globalized world, decisions made in one country often have impacts on people in neighbouring territories. A number of principles for democratic inclusion suggest that voting rights ought to extend beyond the territory of the country in question, at least if decisions are to claim full democratic legitimacy.

***

One popular answer to the ‘boundary problem’ has been that everyone whose interests are affected by a decision ought to have a say over that decision. Since a common justification of democracy is that it gives ‘the people’ power to protect their interests against potentially tyrannical rulers, it seems natural to conclude that all whose interests are at stake should be granted a share of this power. Further, this appears to be an obvious application of the idea that what touches on all should be decided by all. Yet, though it has been suggested that this ‘all affected interests’ principle underlies many existing democratic practices, such as local, regional and indeed national government, it is only imperfectly approximated in existing institutions.

The idea, if fully implemented, would have radical implications, since it suggests that democratic inclusion should not stop at the borders of the nation-state. If a decision has ‘spillover’ effects on outsiders, then they too have a claim to be included in the decision-making. Further, in a thoroughly globalized world, very many of our decisions do impact on people in other countries. Certainly there is a strong case that Mexicans, and many others, should have a say in US presidential elections, since their interests may be substantially affected. But this is only a start; the implications of the all affected interests principle are more radical than this. When the empirical observation that almost anyone may be affected by almost anything is combined with an expansive interpretation of the all affected interests principle, which recommends including anyone possibly affected rather than only those actually affected, it is easy to reach the conclusion that everyone everywhere should have a say in almost everything. Indeed, so radically inclusionary is this principle that it may even require the inclusion of future people and non-human animals.

Whilst advocates of the all affected interests principle are generally untroubled by its seemingly radical implications, suggesting that any reluctance to extend the franchise in these ways is effectively no different in kind from the once widespread reluctance to enfranchise women or racial minorities, others are not so sure, and regard the counter-intuitive implications of the principle as reason to doubt the principle. However, we do not necessarily need something as strong as the all affected interests principle in order to arrive at conclusions that challenge existing practices. Even a more modest principle of inclusion can challenge many existing practices.

***

It is commonly supposed that those who are subject to coercion are owed a justification for this. If democratic inclusion can serve as this justification, this suggests a weaker principle of inclusion, according to which all (and only) those subject to coercion must be included.

This ‘all subjected’ principle is weaker than the all affected interests principle, since its requirements are more modest. Since anyone who is subject to the laws is affected, the all affected interests principle will require the inclusion of anyone who would be included by the all subjected principle. However, not everyone affected by decisions is subject to them, so the all affected interests principle also includes people who would not be included by the all subjected principle. Thus, those who should be included by the all subjected principle are a proper subset of those who should be included by the all affected interests principle, which makes the former easier to satisfy and less radical in its implications.

Nonetheless, while the all subjected principle may be less radically expansionary, it can still have radical implications. For instance, border controls subject outsiders (potential immigrants) to coercion. Thus, if coercion requires democratic legitimation, it requires that ‘outsiders’ be included in decisions over border controls. Where outsiders are not included, border controls lack democratic legitimacy.

As in the case of taxation on the American colonies, someone persuaded by this argument has two potential responses. First, they might decide that the coercion inherent in border controls is unjustified, and hence that states ought to open their borders, doing away with coercion. Second, they might decide that coercive controls require democratic legitimation, and thus that decisions over borders need to be put to a global (or near-global) demos. This second strategy might in practice collapse into something like the first, should this expansive demos reject the use of coercive border controls, but that cannot be assumed a priori. Since an expansive demos might accept some form of coercive border controls, albeit perhaps more limited than those currently in place, these two options remain different in principle, even if they may amount to something similar in practice.

However, it seems that Trump's border wall subjects Mexicans to coercion, even if they are not required to pay for it. Thus, the all subjected principle calls its democratic legitimacy into doubt.

***

Even the all subjected principle is still so demanding that few if any democratic states live up to it in all areas. Again, someone who does not like the implications of the principle can respond by rejecting the principle itself. Perhaps it is simply not true that everyone who is subject to legal coercion must be included in the making of laws. However, something still narrower than the all subjected principle could still challenge existing democratic practices.

Let us return to the idea that there should be no taxation without representation. This proposal requires that all those who pay taxes should be represented. Call this the ‘all who pay’ principle. I take this to be a sufficient, but not necessary, condition for inclusion. That is, I take the principle as one of inclusion only and not also of exclusion; I do not assume that non-taxpayers must be excluded, for there may be other reasons to include them.

As for the other part, I take ‘represented’ here to mean that they must be included in some sense, in particular by being granted the vote (though not necessarily full citizenship rights). While there may be some sense in which person X can represent group Y, without being elected by members of Y, this is less secure than having one's interests represented by someone whom you have authorized and can hold to account. We would surely consider it an affront to democracy if all members of some particular group of competent adults in society were denied the vote, even if representatives were appointed to look to their interests. There would be good reasons to worry about how effective such representatives would be, no matter how good their intentions. Similarly, then, it would be an affront to democracy to suggest that Mexicans’ interests could be reliably served by representatives over whom they had no say.

Many of those who defend the all affected interests principle start from the idea of no taxation without representation. However, they usually generalize from being required to pay to being affected in other ways too. This is well motivated, since otherwise there is a danger that the principle will be misused to exclude non-taxpayers. However, including anyone affected in any way whatsoever seems too broad. It might be thought that anyone who has a preference over the outcome of a decision is thereby affected, even though including them would mean enfranchising nosy busybodies in all kinds of decisions. Therefore, it seems sensible to focus on some particular ways of being affected. Being required to pay taxes is, I think, a particularly compelling ground for inclusion, though probably not the only one.

Being required to pay taxes is narrower than either the all affected or the all subjected principle, since being required to pay tax is to be affected by or subject to only one particular set of laws. Someone could be subject to certain laws, such as traffic regulations, without being required to pay tax. Thus the all subjected principle requires including people who do not pay taxes. But, conversely, anyone required to pay tax is therefore subject to at least certain laws regarding taxes. So, anyone who must be included according to the all who pay principle must also be included by the all subjected principle. And, since the all subjected principle is narrower in scope than the all affected interests principle, the all who pay principle requires the inclusion of only a subset of those covered by the all affected interests principle. Anyone required to pay taxes is affected, but not everyone who is affected pays taxes.

I do not propose to defend the all who pay principle here. Because I have shown that it is less demanding than either the all affected interests principle and the all subjected principle, it should be evident that it is logically entailed by either. This is so because the principle is understood as a sufficient condition for inclusion and not a necessary one. Thus, any argument that everyone affected by a decision (or subject to a decision) should be included in the decision-making entails that all of those who must pay ought to be included, since to be required to pay is to be affected and to be subject to coercion. Given that this principle also enjoys an important place in American history, granting it a further authority, at least in the American context, I take it as not particularly controversial.

To be sure, the imposition of taxes is itself a matter of controversy. There are some, particularly in the libertarian tradition, who object to taxation altogether. However, I am not saying that the state should or should not tax people; that is something that I take no stance on here. What I am saying is that if the state taxes people, it ought to include those who are taxed in its processes of collective decision-making. This principle can be endorsed, even by those who oppose taxation. Yet this principle implies that, if Mexicans are to pay for Trump's border wall, then they must be included in the demos.

I do not address arguments against taxation here. My claim is that the state has no right to tax anyone that it excludes from the demos. The libertarian may wish to go further, and to say that the state has no right to tax anyone, period. I do not wish to commit myself to that, but, since it entails my own more limited principle, I have no need to argue against it.

There are different ways in which this inclusion could operate. One possibility would be to have included Mexicans in the US presidential election. Had they voted in significant numbers, this might plausibly have led to a victory for Hillary Clinton, rather than Trump. However, since the presidential election had important implications for many other domestic policies, such as American healthcare reform, this would also have resulted in Mexicans being included in decision-making on many other matters that they had no claim to inclusion in. This is not necessarily objectionable on the principle propounded here, since I have only proposed taxpaying as a sufficient and not a necessary condition for inclusion. Nonetheless, it could be seen to reduce the self-determination of the American people over matters that affect only them. However, the all who pay principle only requires that Mexicans be included in decisions over the border wall, and anything else that they are to pay for, not in the presidential election itself. Thus, an alternative possibility would be for the presidential election to take place as normal, but for Trump's border wall to need the approval of the Mexicans through some further democratic process. This, in turn, might have been a direct democratic process (like a referendum) or something involving Mexican representatives.

***

I have argued that democratic principles require the inclusion of Mexicans in any decision to build a US–Mexico border wall, if they will be required to pay for that wall. But it might be objected that this is incoherent, for we cannot know whether Mexicans will be required to pay for the wall unless we have already determined whether the wall is to be built. If the decision is to build a wall, and demand that Mexicans pay for it, then the Mexicans ought to be included, or have been included, but it is now too late to include them. Alternatively, if they are included in the decision, then the decision may be not to build a wall in the first place (or to pay for it with American money), in which case the Mexicans had no claim to inclusion. Since we cannot identify what must be paid for until a decision has been reached, we cannot identify who must pay until after the decision, and therefore it is impossible to have those who will pay make the decision.

The problem here mirrors a familiar problem with the all affected interests principle. It has been argued that we cannot identify who is actually affected by a decision prior to it being decided, as this would be incoherent. However, we can distinguish being affected by an outcome and being affected by a choice. It is true that we cannot identify those affected by an outcome without knowing what the outcome is, but we can identify those who may be affected by the different outcomes that are being chosen between. These people, then, are affected by the choice. It does not matter whether the people in question actually pay or not, what matters is who would be required to pay for the decision in question. If the choice confronting us is whether to make Mexicans pay for a wall, then they ought to be included, because they will be asked to pay if the decision is yes.

One possible response here is to deny that Mexicans will pay for the wall. Of course, if Mexicans do not actually pay for the wall, then the all who pay principle does not require their inclusion. The American people can unilaterally decide to build a wall along their border if they are paying for it. But I am taking Trump at his word here when he says that the Mexicans will pay. I think this is probably unrealistic and it is possible that we should not take Trump's claims at face value. Perhaps his pledge to build a wall is best understood metaphorically, as an indication that he will be tough on immigration. However, as far as I am aware, Trump has said nothing to suggest that his words are not meant literally. Whether or not a wall is actually built, and whether or not Mexicans are made to pay for it, this seems to be what he promised. I am interested in the implications of this promise if it is taken literally, not whether it should be taken literally.

One way in which Trump has suggested funding the wall is through a tariff on Mexican imports. This, supposedly, would amount to Mexican exporters paying for the wall. However, this is a misunderstanding. The result of such a tariff will be that Mexicans sell fewer goods to the United States at higher prices. We can see the result as the cost of the tariff being split between Mexican exporters and US consumers, with this split depending on what economists call the price elasticity of demand. To the extent that US consumers end up paying higher prices, they are actually footing some of the bill. Thus, it is unclear whether this particular policy lives up to Trump's claim that Mexicans (and Mexicans alone) will pay for the wall. However, in so far as Mexicans still pay something, the all who pay principle still requires their inclusion, even if they do not pay everything.

At this point, it might be objected that the all who pay principle must be mistaken, for it is too demanding when we consider everyone who pays taxes. At least in a state that imposes sales taxes on many goods, some taxes will be paid by children and transient visitors whenever they purchase anything that is subject to a sales tax. The all who pay principle would therefore require the inclusion of both children and visitors who pay taxes. If this seems over-inclusive, then it may be taken to show that the all who pay principle is mistaken.

This is an important challenge to the all who pay principle, but it must be remembered that a similar challenge also applies to both the all affected interests principle and the all subjected principle, which also seem to require the inclusion of children and transient visitors. If a suitable answer can be found in these cases, then the all who pays principle can presumably avail itself of the same response.

Another solution would be to finesse the all who pays principle somewhat. Perhaps it is too demanding to suppose that anyone who pays anything must be included in the demos. The principle might be refined, so that it covers only those who pay at least a certain amount or in a certain way, such as through direct taxation. Alternatively, perhaps, it may be interpreted to cover only those who pay in a regular, ongoing fashion. The transient holidaymaker may pay a small amount of tax, during a visit of a week or two, but does not expect to continue contributing towards future projects. If we seek a reasonably stable demos, then we may wish to focus on those who have either a reasonably lengthy history of paying tax or who reasonably expect to continue doing so for some time. Since only these people have a long-term stake in the society, perhaps only they have a claim to inclusion.

It might be replied that, if we reinterpret the all who pay principle in this long-term manner, it casts doubt on whether the Mexicans ought to be included in decisions over the border wall. To be sure, taking a more restrictive interpretation of the principle will reduce its demands for inclusion, but it may still require the inclusion of Mexicans. First, the border wall would be an expensive undertaking, so its cost would presumably be spread over a considerable period of time. Second, though building the wall is a one-off cost, there would be ongoing costs for maintenance, security, and so on. Assuming that these costs were also to fall on Mexicans, being ongoing costs of the wall, then Mexicans would be paying in perpetuity and this would give them a claim to inclusion, even on a narrower interpretation of the all who pays principle.

***

Trump's pledge to make Mexicans pay for his proposed border wall provides a clear illustration of the importance of the democratic boundary problem. To be sure, it is not the only one. The UK's 2016 referendum on exiting the European Union (‘Brexit’) also offers another topical example. However, Trump's border wall case is particularly salient given the resonance of the idea that there should be ‘no taxation without representation’ in American history. Since this was one of the founding principles of the American republic, Trump's apparent violation of this principle with regard to Mexicans is a particularly egregious wrong.

Further, I have shown that the principle that those who pay should be included in decisions as to how ‘their’ money is spent need not depend on some more radical principle, such as the all affected interests principle or the all subjected principle. Though the all who pay principle is entailed by either of these popular alternatives, we might accept the all who pay principle without committing ourselves to these further principles. While the all who pay principle leaves open the possibility that there are other grounds for inclusion, it can be seen as a point of overlapping consensus between the main contenders in the debate.