How rational can we expect nation-states to be in setting immigration policy?

A recent comment by Christopher Chang made a point that I feel we haven’t adequately addressed on this site: if open borders is such a great deal, why has nobody tried it? Using a similar structure of argument as Bryan Caplan employed in his blog post on motivating sheep or Gary Becker employed in the now-standard argument that market forces gradually erode discrimination, Chang writes:

It does not matter if 90% of actors are inefficiently biased, as long as some subset of the other 10% knows what they are doing; that subset, and its future imitators, win in the long run. This has happened over and over and over again in political, economic, and military history. Both you and Paul fail to comprehend the nature of “collective assessment” that requires only one yes vote. As I’ve mentioned numerous times, you can even provide much of the yes vote yourselves, and according to all your economic claims, this would be quite lucrative. (This does not mean that any particular individual’s failure to do so is damning; life circumstances frequently interfere. But the fact that none of you have done anything like this does, in fact, add up to a strong revealed preference against your claims. Your failure to even seriously openly discuss this among yourselves strongly implies that you do not actually want your rosy projections tested.)

I like the general reasoning behind the generation of these types of questions. Claims that the status quo is radically suboptimal should be met with skepticism, and should be taken seriously only if the people making the claims can offer a convincing explanation for why the suboptimal system is stable. In this post, I explore some possible explanations. Read and judge their strength for yourself!

Note that this post is not, in and of itself, a complete argument for open borders, or open borders with keyhole solutions, being optimal. It’s rather an explanation for how the current state of the world (far from open borders) is consistent with the possibility of open borders being a lot better. It seeks to address a potential inconsistency, rather than offer complete proof. Therefore, my rhetoric in the post will assume the open borders position and simply demonstrate that there are no obvious contradictions.

There’s a connection between this post and my earlier post on whether migration levels under open borders would be optimal, too high, or too low. But whereas that post is about the decisions of individuals holding state policies constant, this post is about the creation and tweaking of the policies themselves.

#1: The gains from pure open borders go to quite an extent to migrants and their descendants, and even though existing residents of migrant-receiving countries gain somewhat, the gains are a lot less

Why don’t we have pure open borders, if it benefits the world so much? The short answer is that the people with the power to decide this (the people in political power and the voters and special interests that they cater to) are not the people who benefit the most from open borders. As Nathan Smith notes here and here, immigration policy is quite “undemocratic” in the sense that potential immigrants have no (direct) electoral say in a matter that affects their freedom.

This point has a number of different aspects:

  • If the gains from migration went mostly to the immigrants and not to non-immigrants in the target country, the people who gain the most don’t have a say in the electoral process. Note that this point is valid even in the absence of disparity or asymmetry between nations. If immigration from Canada to the US significantly enriches the Canadians who migrate to the US, but has little effect on US natives, then US natives (who vote in the elections) have little incentive to push for freer migration from Canada.
  • One possible remedy to the above would be to push for free migration through reciprocity, for instance, freeing migration from Canada to the US in exchange for freeing migration from the US to Canada. This would work in the case that the fraction of the population in either country that has an interest in the option of migrating is large enough: if enough Americans want the freedom of easy migration to Canada, they may vote for a treaty that frees migration both ways. If, however, the fraction of people interested in migrating is small, then they may not be able to push for freer migration even if the absolute gains they experience are huge. That’s because democracy is based on counting votes, rather than on winners compensating losers.
  • In the current world, there are significant international disparities in wealth and wages, and a strong directionality to potential migrant flows. In light of this, the spotlight falls on the migration policy of countries that have greater per capita income or wages or are otherwise attractive migrant destinations. The degree of solidarity between potential migrants and the set of people who have the most influence over the most relevant migration policy is now quite low: the former are people from low-income countries, and the latter are people of prosperous countries. Even though the latter set is expected to gain somewhat in expectation, the gains are smaller in absolute terms, and even smaller when viewed as a proportion of how well off they currently are.

Note that many of these can be fixed, at least in principle, if we relax from pure open borders to open borders with appropriate keyhole solutions such as pro-native tax-and-transfer schemes. So why don’t we have instances of the latter? Actually, we do, to some extent. We’ll get to this later in the post.

#2: The full gains from open borders take time to materialize

Estimates of significant increases in world production after open borders are not estimates of overnight gains. Rather, these are estimates of how the world would look a decade or two from the opening of borders, relative to how it might look in the counterfactual. Even if the estimates were correct, the full magnitude of the gains would be felt after a fairly long time-lapse. Thus, there may not be good electoral incentives for democratic governments to support freer migration for the economic benefits. Some of the economic benefits would be reaped immediately, but it would be an order of magnitude less than the long-term gains. Note that this is similar to the reason why we expect individual migration to be less than what seems economically optimal, as discussed in my other post.

#3: In so far as there is a citizenist case for open borders, it relies on a tax-and-transfer scheme combined with somewhat draconian enforcement

Nathan Smith has previously argued that there is a citizenist case for open borders: use a tax-and-transfer scheme such as immigration tariffs or DRITI to hold natives harmless and distribute the gains away from migrants and towards natives.

Schemes like DRITI present a dark side: they exacerbate the visibility of poverty and, even as they reduce the unfairness of the system as a whole, make it more visible. So, in addition to open borders advocates who’d worry about such schemes (see, for instance, here and here), there are many others, such as immigrant rights activists, who would reject these schemes. Even if open borders started out with such “keyhole solutions” it’s not clear that they’d be stable.

The combination of citizenism and the form of local inequality aversion make tax-and-transfer-based keyhole solutions a hard sell in many countries. I don’t think the problems are insurmountable. But I don’t think it’s surprising that we haven’t seen a lot of progress on these so far. It’s fruit, but it’s not very low-hanging fruit.

Incidentally, one prediction of this setup is that countries that are less democratic, and where governments have more authority to carry out more draconian enforcement, might be more likely to have implemented citizenistic migration liberalization. This is indeed the case, as we’ll see in #5.

#4: The incentives of democracy

The incentives of the politicians and bureaucrats running the government are not perfectly aligned with the long-term interests of the citizens. There are two broad problems:

  1. Citizens themselves don’t know what’s best for them, so they may not reward their representatives in government for choosing better policies or delivering better outcomes.
  2. Politicians often cater to special interests rather than the needs of citizens.

In the context of why there hasn’t been more significant migration liberalization, I expect (1) to be a far bigger reason than (2). The general phenomenon of political ignorance in the electorate in advanced democracies has been well-studied (the political knowledge elsewhere is highly unlikely to be better, and quite likely to be worse). Explanations such as rational ignorance and rational irrationality have been offered. Given generally low levels of political knowledge and decision-making skill on the part of the electorate, we shouldn’t have strong reason to expect a democracy to converge to a good outcome. But we might still expect that, by random chance, some democracies would converge to good outcomes in a given area. So a bit more is needed.

Bryan Caplan has posited that one particular aspect to voter irrationality is anti-foreign bias: people systematically underestimate the benefits of interactions with foreigners. Are people inherently anti-foreign? I think that people have inherent tendencies to support their ingroup and resent or discount the welfare of outgroups. But the particular use of nationality as the criterion to define ingroup and outgroup is probably an artifact of the political process: there is an existing governance infrastructure that facilitates discrimination on the basis of nationality, and an existing ideological infrastructure that gives particular importance to national identity. So people’s diverse ingroup-outgroup choices get projected to divisions based on nationality and citizenship even if that’s not the best way of describing the distinctions in their own minds. The upshot is that we might expect the political process to produce an anti-migration bias relative to what’s optimal, and while this reflects some discomfort that individuals experience interacting with foreigners, it’s often a result of a political process translating other forms of discomfort people have to the language of discrimination based on nationality (this is a somewhat tricky point, and I hope to elaborate in a future post).

#5: Some countries that are undemocratic or less democratic, and have more deference for elites, tend to have citizenistic and numerically liberal migration policies

Carl Shulman has been doing yeoman’s work of late examining some interesting “open borders with keyhole solutions”-type regimes. He looked at Singapore in a blog post titled Migration levies and unskilled labor mobility in Singapore, where he discussed Singapore’s large temporary guest worker program for low-skilled migrants, that combines fairly huge numbers of migration with fairly stringent restrictions on what migrants can do, as well as taxes on the migrants that make them fiscally good for the government. In a blog post titled What does migration to the United Arab Emirates tell us about labor mobility?, he looked at the UAE, where a significant majority of the population is foreign-born, and where there are significant differences between the rights and privileges accorded to a native elite and a large foreign-born workforce. In both cases, their policies created a win-win for natives, the government officials, and migrants. In Singapore, the success of these policies arises from an effective one-party system and considerable deference to elites in policymaking, despite the general population being less pro-migration than in many First World countries with far stricter limits on migration. The UAE is a federation of hereditary monarchies, which insulates it from the pressures of competitive democracy. The population is also less likely to push for liberal ideals of equality that jeopardize the stability of keyhole solutions.

Relatedly, a recent book called The Price of Rights by Martin Ruhs (to be reviewed later) empirically came to the conclusion that there’s a trade-off between the number of migrants a country admits and the package of rights that are accorded to people after they migrate. A similar point had been made in a post by Michael Carey a while back on immigration and class struggle. If modern liberal democracies tend to be strong on the package of rights and privileges, this (often) comes at the expense of the number of migrants they admit.

#6: If many countries tried citizenistic open borders, competition would drive down tariffs eventually, bringing the world closer to open borders

In #1, we (sort of) ruled out the plausibility of pure open borders (barring dramatic changes in people’s views of moral permissibility and side-constraints). But in #5, we pointed out that we do have partial approximations to “open borders with keyhole solutions” and these could be taken further. So how far are “open borders with keyhole solutions” from pure open borders?

We can think of pure open borders as migration with zero government-imposed barriers or taxes. Suppose most countries have prohibitively high barriers, and a few countries experiment with selectively reducing taxes to the level of “open borders with keyhole solutions.” These few countries can still afford to keep the taxes at their profit-maximizing level, because they’re effective monopolists: all the other countries are out of the running because their barriers are too high. The policies of these few countries are a Pareto improvement over the status quo, but they still carry the inefficiencies of a monopoly. If, however, more countries start getting drawn into the game of open borders with keyhole solutions, then there is more competition between countries that exerts a downward pressure on the taxes and barriers. Thus, as more countries do it, we get closer to open borders. We probably don’t get anywhere near pure open borders, but we do get a lot closer than if only one or two countries were trying it out.

PS: Open borders isn’t the only policy proposal for which we can ask this sort of question. I’m quite curious to hear the thoughts of proponents of drug legalization, free trade, organ trading, and other such cutting-edge proposals. The situation with some of these seems to be a bit better than for open borders, but not by a huge margin. For instance, consider the case of drug legalization. According to this Wikipedia page, there is only one country, Uruguay, where the possession, sale, transport, and cultivation of cannabis (marijuana) are all legal. But there are a number of nation-states (plus member states in nations) where the possession of marijuana is legal, or illegal but decriminalized, and there are others where marijuana use is de facto tolerated. So even though full-blown legalization is rarely embraced, we have enough variation in the direction of legalization to address the question of “if it’s such low-hanging fruit, why has nobody plucked it?” This is roughly similar to the situation with open borders.

PS2: I wrote up a condensed version of an early draft of this post in the form of an Open Borders Action Group post.

One thought on “How rational can we expect nation-states to be in setting immigration policy?”

Leave a Reply