Immigration is arguably the most heated issue in global politics at present, not only at the center of the 2024 US presidential election but also consistently reported to be one of the dominant areas of concern by electorates across the West. Much of the public discourse on the topic proceeds along well-worn paths of cultural and economic disputation. Is immigration harming public safety? Are newcomers assimilating successfully? Does immigration depress the wages of workers, or encourage economic growth, or strain public services, or invigorate key industries? While the pitch at which they are now debated is striking, these questions are longstanding features of Western politics.
However, one theme has recently achieved a new level of prominence in our immigration contretemps. Unlike past versions of the debate, at least in the US context, current discussions have paid more attention than hitherto to a specific political concern that is distinct from the familiar cultural and economic ones. The concern is that immigration is undermining authentic democratic self-rule. The most famous spokesperson for this view is none other than the richest man in the world (himself an immigrant, as it happens).
On Sept. 14, Musk bemoaned the “massive voter importation to make swing states permanently blue and turn America into a one-party state.” Shortly before, in response to a Nancy Pelosi interview in which she spoke in favor of a pathway to citizenship for a large number of undocumented immigrants, Musk had lamented that “the Democratic Party is importing voters. It’s as simple as that.” Musk is, naturally, far from alone in sounding this alarm, even if he is the most famous person singing from this hymnal. The critic Jacob Siegel warned that “what may turn out to be the most consequential of the ‘targeted policies’ that the party-state uses to engineer elections involves immigration.” Musk expanded on this line of thought in an essay-length tweet a few weeks ago, in which he turned the “threat to democracy” charge beloved of Trump critics back at Harris on the grounds that her administration would carry out a program of naturalizing illegal immigrants which would ensure permanent Democratic rule going forward.
Whatever one thinks about the ideological valence of these messages or the plausibility of their prognostications, the newfound elevation of this framing of immigration as affecting the integrity of democracy cannot be lightly dismissed. Indeed, wherever one comes down on the substantive issue of what the best immigration policy should be, the currents of discontent which Musk incarnates raise an interesting problem for political and constitutional theory. Namely, what kind of issue is immigration, and what sorts of institutions are appropriate for deciding immigration policy?
We often speak of the system of government that prevails in the West today as “representative democracy.” But the fact is that the ideal type of this regime, in which elected officials pass legislation and run the government while what James Madison called “the people in their collective capacity” are totally excluded, is scarcely to be found. Liberal states in the modern world tend to contain a combination of direct popular input and more traditional representative institutions. Only a handful of existing democracies have never held a national referendum. And although the United States is one of those few, it has nonetheless been a leader in “direct democracy” with innumerable ballot measures at the local level and various styles of initiative and referendum operative at the state level. (For instance, of the 50 states, Delaware is the only one that does not require a popular vote to amend its constitution). Indeed, direct democracy has been present since the birth of the Republic, with the Massachusetts Constitution of 1780 being the first constitution in history to be ratified by a popular referendum. And the US constitutional amendment procedure, while hardly a straightforward national vote such as has been on display in recent high-profile contests in Ireland, was meant in its very baroqueness to guarantee that changes to the basic governmental structure of the country would demonstrate a greater degree of popular legitimacy than a regular law passed by Congress.
This mixed character of modern democracy—largely representative, but with heavy doses of direct popular involvement—naturally raises the question: What is the appropriate division of labor? When should referendums be used, rather than the ordinary legislative process?
“Referendums should be held wherever the normal relationship of choice between legislators and the public is flipped.”
One important answer that has been given—for instance by the highly influential political philosopher Dennis Thompson—is that referendums should be held wherever the normal relationship of choice between legislators and the public is flipped, so to speak. Voters are supposed to select their representatives; in more technical verbiage, the public is the principal, and the politicians are their agents. However, there is a class of issues in which it looks like the direction of travel has been reversed, where the agents are in a way choosing their principals. This class encompasses decisions which affect the electoral system or the shape of the sovereign power. To allow legislators to change the electoral system—how seats are apportioned, how votes are aggregated, how districts are drawn—would be to allow them, it seems, to “choose their own constituents.” Likewise, decisions that fundamentally alter the nature of sovereignty—by joining a supranational political body like the European Union, or seceding from an existing state, or transforming from a unitary to a federal republic, or merging two polities into a single one—require a direct-democratic verdict because otherwise the politicians would be deciding for themselves the character of the sovereign they are supposed to be serving. Instead of acting on behalf of the sovereign people whose consent to the constitutional system, according to a democratic outlook, alone legitimizes the power of the existing lawmaking processes, a legislature that could pass legislation on its own to, say, force a province out of the union would be shaping the sovereign after its own fashion. In the language of political theory, this would amount to a kind of usurpation of democratic authority.
Nor are theories like Thompson’s recent innovations. Since the Age of Revolutions, leading thinkers of the modern republic have argued that when it came to constitutional issues—by which they meant above all those issues which Thompson emphasized, to do with the organization of the electorate and the fundamental distribution of political power—ordinary political bodies and officials should not be able to implement changes on their own. As the great theorist and participant in both the American and French Revolutions, Thomas Paine, wrote, “it is not for the benefit of those who exercise the powers of government, that constitutions, and the governments issuing from them, are established,” and so where questions of the basic structure of authority were in play, “the right of judging and acting” returned to the citizenry. Likewise an archetypal figure of the Enlightenment and author of one of the most radically democratic constitutions ever written, the Marquis de Condorcet, argued that to permit the legislature to implement such changes was “dangerous and absurd,” allowing it to “become independent” of the constitution and the people whose sovereignty was supposed to be instantiated by it. Subsequent practice has endorsed these insights, and to a large degree Thompson’s own theorization was not a philosophical novelty but a rationalization of what liberal democracies have come to treat as best practice. With the explosion of referendums since World War II, it is precisely those areas which Thompson identified that furnish the most frequent subjects for this mechanism: proposals to alter the electoral system or the distribution of sovereign power make up a large portion of the total national-level referendums that have been held worldwide.
How does this historical and theoretical excursus relate to our original subject? I would suggest that the logic about when to use referendums articulated by Thompson, Paine, Condorcet, and accepted by the numerous states which have made this their practice, ought to be extended to immigration. For whatever one thinks about the substance of the issue—whether one thinks immigration is a net-positive or net-negative, or whether some kinds of immigration are the one and some the other—it is an indubitable fact that policies around immigration, naturalization, and pathways to citizenship do greatly affect the nature of the sovereign people. Indeed, the questions of whether two adjoining countries should unite, and of whether one of these two countries should permit large-scale immigration from the other, pose many of the same issues regarding the nature of the democratic people who is supposed to be enjoying the right of collective self-determination. Immigration and naturalization policy determine who will join the rank of citizen, and therefore, in a democratic country, who will participate in the exercise of popular sovereignty.
“This is to permit politicians a say in deciding over whom they will rule.”
For a kindred reason that an ordinary legislature or executive should not be able on its own to cede a portion of policymaking power to a supernational body, or to eject one of its territories from the union without the consent of that portion of its citizens, or to replace a bicameral with a unicameral legislative power, the ordinary legislature or executive should not have discretion to set immigration and naturalization rules on their own, absent direct-democratic recourse. For this is to permit politicians a say in deciding over whom they will rule. And this stands in contravention of the normative democratic direction of authorization, according to which the people, which is sovereign, selects its government, rather than the government selecting its people.
Immigration, in other words, should be treated as belonging to the realm of constitutional or fundamental law, and thus decided directly by the people, rather than to the realm of ordinary law, where our normal policymakers’ say is determinative. This is the truth implicit in the discontent that Musk and others are expressing, which one should recognize even if one disagrees with their assessment of the social-economic consequences of immigration. There is indeed a democratic deficit in leaving the foundational question of who our fellow citizens will be to the mere domain of party conflict, where any number of other considerations must necessarily enter into our decisions and the issues are muddled with judgments of personalities.
Note the exact character of the argument here. Its crux is the effect that immigration has on the composition of the citizenry, which in a democracy is meant to be sovereign. It has nothing to do with a verdict that immigration is more important than other issues, like taxation or crime or antitrust or healthcare. What dictates that a referendum is the fit instrument here is not momentousness of impact, but the fact that who the people is is in play. The situation is perfectly analogous to the issues that democratic theorists already see as requiring direct democracy. Switching to a proportional representation system or accepting the jurisdiction of the European Court of Human Rights might have much less effect on society all things considered than passing a Green New Deal-style legislative package or declaring war, but the former are nonetheless issues that require direct-democratic consideration.
Apart from this basic normative case for why immigration is not an appropriate issue to be left to ordinary governing institutions, there would appear to be ample prudential grounds for those across the ideological spectrum to welcome this conclusion. For the fact is that the public in just about every single Western democracy has wanted reduced levels of immigration for some time now, and it is the slowness or unwillingness of the regular political system to respond to this demand that seems above all to have been the accelerant of the populism feared by so many. In those countries that have taken meaningful steps to restrict immigration, like Denmark, conventional center or center-left parties are flourishing. Had immigration been treated as a matter to be decided directly by the people to begin with, it is at least plausible that this bête noire of liberals everywhere would never have gained momentum. If populism is fueled by distrust of traditional institutions like parties and assemblies, then making it impossible to accuse the latter of disregarding the popular will in order to remake the citizenry as they see fit by simply taking this most flammable issue out of their hands looks like a promising avenue for restoring confidence.
Nor need advocates for a permissive liberalization regime despair—they would simply have to reaccustom themselves to persuading their fellow citizens, rather than relying on the consistent gap in support for high levels of immigration between the political class and average citizens to accomplish their goals. In the United States, the overturning of Roe v. Wade removed the question of abortion from the judiciary, but it also incidentally removed it from the arena of party conflict in a number of states, which put the question directly before the electorate. In those cases, including a few states which were deep red and where it had been presumed the pro-life position was ascendant, the left-liberal position on abortion won at the ballot box. It is hard to predict which arguments and positions will prove most resonant when issues are moved from one domain to another (from courts to legislatures, or from either to referendums).
In many Western countries, adopting the referendum as the instrument for settling immigration and naturalization policy, whether through law or custom, would be fairly straightforward. Even in the United Kingdom, historically the land of parliamentary supremacy par excellence, a norm has been established of using referendums where the fundamentals of sovereignty or peoplehood are at stake. In many places, installing immigration referendums merely requires recognizing the kinship between immigration and the other subjects for which referendums have come in for common use.
In the United States, our peculiar constitutional structure makes setting up any such process very hard to imagine. Nevertheless, it is still useful to understand the ideal, so that we might attempt to approximate it with the means we have available. A norm could be established whereby politicians would treat immigration as belonging to the realm of constitutional politics, as it were, and refuse to pass policy changes via narrow partisan majorities or use the agencies of the executive branch to substantially alter policy on this front. A bipartisan commission could be established to write model legislation after a series of national polls had ascertained the wishes of the people on the subject—legislation which the leaders of both parties could commit in advance to accepting. Despite all the difficulties of amending the US Constitution in a polarized age, it is not unthinkable that a constitutional amendment setting new terms for immigration policy might come about. Any number of modalities which might do some justice to the recognition of the intrinsically direct-democratic character of immigration as an issue could potentially arise. The most important thing is to see that it is politically and constitutionally misguided to continue to treat this issue as one of ordinary partisan contestation or regular governmental regulation.
Arguing in favor of nationalist independence movements in the mid-19th century, the greatest liberal philosopher of all, John Stuart Mill, wrote that “one hardly knows what any division of the human race should be free to do, if not to determine, with which of the various collective bodies of human beings they choose to associate themselves.” The argument would seem to apply to immigration as to secession or state-building. If the people are not fit to determine who will become their fellow-citizens, why should we think they are fit to do anything at all?