Immigration is complicated

Author’s note: this is a non-public page, deliberately not linked from the main blog. It’s fine to share the link with people but please use good judgment. This is currently in bullet point form. I might gradually write it up into an essay.

Epistemic status: low conviction



Millions of Syrian refugees fleeing their country. Mexican toddlers separated from their families at the United States border. Migrants languishing in refugee camps in Manus and Nauru. Canada’s positive attitudes to migrants despite high immigration rates. Denmark’s controversial new assimilation policy. Three quarters of Silicon Valley’s tech workers are immigrants.

Immigration is a complex and sensitive subject and many people have strong views about it. This doesn’t mean that it must be politicized – love is also complex and sensitive and people have strong views about it, but it’s not political.


Nationalists underestimate the economic benefits. Progressives underestimate the social costs. Classical liberals/libertarians overestimate the economic benefits. All of the common value systems tend to underestimate the importance of either economic effects, social effects, or both.

Compared with other divisive topics, immigration has an extra obstacle: the economic effects are closely tied to the social effects. We are forced to understand both of them if we want to have a substantive discussion.

The following discussion is from the viewpoint of the host country, because the host country is the one with the power to set immigration policy. Immigrants have no negotiating leverage, and the home country has very limited power to control emigration. [TODO: Definition of what immigration is – different types of immigration][i]



  • Less informed people often think that immigrants are negative for the host country. They compete for jobs and drive down wages, because increasing the supply of labor will reduce its price. Others fear that immigrants will enjoy benefits without contributing to the host country.
  • The informed layperson generally assumes that immigrants are a net positive for the economy. They drive innovation and entrepreneurship. Some (wrongly) believe that immigrants can help the population growth rate in low fertility countries.
  • Another argument is that the benefits to the host country are distributed unequally, with some areas gaining more than others, and some people gaining more than others. How big of an issue this is depends on how much you think inequality matters.
  • This more informed view is generally true, but there are several major caveats:
    • Not all immigration is economically beneficial. There are lots of statistics about how immigrants tend to be well educated, work harder, high birthrates, and likely to start new businesses, but this is a result of strong selection effects. Increased immigration will, ceteris paribus, lower the quality of incoming immigrants.
    • A rudimentary mistake is seeing that highly selected immigrants are beneficial, and concluding that much more immigration would be beneficial. This assumes that immigrant quality doesn’t change as immigration increases.
    • A more common mistake is assuming that immigrant quality declines gradually. The implication is that the number of foreigners wanting to immigrate rises linearly as the ease of immigration increases, but there is no evidence for this.
    • The policy of ‘fully open borders’ is an extreme example of this: if implemented, it would knock the foundation out from what makes the statistic true.
    • The costs to the migrant are often underestimated. For example, some research suggests that the median moving cost within the United States is around $300,000, based on people who don’t relocate even when the economic benefits are large.[ii] The costs to the marginal migrant is significantly smaller, but migration across borders is more costly than within a country.
    • The host country, if it is of a reasonable size, is not monolithic. Some areas may benefit, others may suffer.
  • In conclusion, there are probably economic benefits to both the immigrant and the host country, but these benefits are not uniform and the change in the economic returns with increasing amounts of immigration is unknown.


Social capital

  • Social capital is not about specific races and demographic groups being inferior or superior, but about the cultural distance between groups that must coexist and sustain institutions. If the cultural distance is too far, the basic level of trust necessary to, say, maintain a functioning government doesn’t exist.
  • Social capital is difficult to quantify, because it’s based on the configuration of personal and societal relationships, and is not fungible.
  • Useful ideas about social capital can come from sociology, political science, history, anthropology, and other humanities fields where quantitative methods are less important and qualitative methodology is comparatively more important. These fields are less prestigious than economics because of weaker methodology and less emphasis on quantitative methods, but their area of study is more relevant.
  • Attempts to study social capital based on an economics perspective are ineffective, because of the lack of quantitative data and difficulty in classifying the data means that social capital is not amenable to quantitative methods. Describing patterns is different from creating hypotheses.
  • Unfortunately, because of the interplay of social capital and economic benefits, we can’t afford to ignore social capital.
  • Social capital is underrated by all common political persuasions except for social conservatives. There are four likely reasons for this:
    • Since it’s harder to quantify, it’s lower status in the academic world.
    • People are reluctant to talk about social capital, because they fear being associated with racism.
    • It’s hard to intuitively grasp how large an impact social capital has, if you haven’t lived in several places with different levels of social capital for significant amounts of time. Like a fish not noticing the water it’s in, we don’t notice how much social capital our society has, unless it changes suddenly. Since the majority of people don’t have this experience, it leads to social capital being underestimated in its importance.
    • It’s also a dynamic produced by complex networks of institutional and social interactions. This makes it hard to “see” how social capital works; we can only notice its effects. Sometimes, its effects are misattributed to other causes.
  • The dictatorship of the small minority concept – the potential danger that a tiny but highly motivated and cohesive minority can compel the majority to accept its demands – is seldom considered.
  • Social capital is even more likely than economics to have non-linear changes in the payoffs from immigration, due to groups having feedback loops and network effects. Tipping points are not visible until they have been crossed.
  • In conclusion, there are probably social costs to the immigrant and the host country, but these costs are not uniform and the change in social capital with increasing immigration is unknown.



  • Assimilation is the process in which a minority group or culture comes to resemble those of a dominant group. Social scientists rely on four primary benchmarks to assess immigrant assimilation: socioeconomic status, geographic distribution, second language attainment, and intermarriage.
  • Assimilation is important because it reduces the social cost of immigration, and increases the economic benefits of immigration. It does this by increasing the amount of common knowledge, thereby reducing social frictions and increasing societal-level trust.
  • Assimilation may also benefit the host country, depending on which elements of the immigrant’s culture gets absorbed into the host country’s general culture.
  • Adults usually do not assimilate beyond very practical things that they can’t compromise on, like following the laws of the host country. Assimilation is mostly about children, and has to do with the degree of social and economic independence of the recent-immigrant ghettos.
  • If the ghettos are very independent, children will grow up with minimal contact with non-ghetto people. In France, for example, predominantly Muslim ghettos have minimal contact with other places, likely a consequence of its 1974 decision to change its immigration policy from assimilation to multiculturalism.
  • If the ghettos are more connected to the rest of society, they become transitional places where recent immigrants live in a semi-cushioned cultural half-way house. Their children are fully-conversant in the broader culture and typically move out of the ghetto. In New York City’s Chinatown, people from everywhere else regularly visit and do business, shop, eat, and so on.
  • The mechanism for these results is that children take on the culture of their peer group, not of their parents. If social institutions are set up such that peer groups are mixed or majority-host-native, the tendency will be for the immigrant child to assimilate strongly. But if the institutions are separate for the immigrant culture, and the peer groups of the children are self-sustainably foreign-cultured, then assimilation will not happen.
  • It’s important to encourage immigrants to leave their ghettos and familiarize themselves with the host country’s culture. We must avoid the permanent ghettoization of immigrants, which prevents the kind of social interchange which can integrate the foreigners into the fabric of broader society and allow the foreigner to create social capital within the broader society.


Factors affecting assimilation

The social institutions of the host country is by far the most important factor in determining the ease of assimilation and how much immigration is optimal, but there are other factors too. Here is a list of the other important factors that affect both economic and social outcomes. Most of these factors are under-discussed in popular discourse, and many are non-obvious.

  • The host country’s attitude towards immigration: multiculturalism vs assimilation/melting pot.
    • Multiculturalism is the recognition that different cultures have value, and that immigrants should not be forced to accept the cultural views of the host country; diversity creates value. In multiculturalism, immigrants are allowed to keep their distinct cultural identity. Assimilation, or “melting pot”, is an approach which includes the immigrants’ culture into the broader social culture. Elements of the immigrant’s culture are accepted, but in return, the immigrant should take on the host country’s culture.
    • Multiculturalism reduces the social costs to the immigrant and increases the social costs to the host country.[iii] But this is a sliding scale rather than a binary; few countries practice complete multiculturalism, and countries vary widely in the amount of assimilation they expect of immigrants.
    • Many countries claim to be multicultural, but the de facto approach is much closer to assimilation/melting pot.
  • Assimilation power of host country. Countries vary in their capacity to assimilate immigrants. Countries that are generally admired by immigrants will have an easier time of it, and it is helpful if the culture is reasonably cohesive. For example, the United States has high assimilation power, possibly because it has high cultural power.
  • Required assimilation level of host country. If the host country requires a very high degree of assimilation, then immigration is more costly. If the immigrant can assimilate while making fewer changes, then immigration is less costly. Examples:
    • Japan is high. Therefore, although Japan has [high soft power], we would expect that they would have low amounts of immigration. Who wants to deal with stuff like this if they don’t have to?
    • China is high. The Chinese national view sees the world as separated into “civilized people”, which refer to ethnic Chinese, “semi-civilized people”, who are on the periphery and in nearby countries, and “barbarians”, which are seen as incapable of assimilating. This mindset leads to low levels of assimilation.
    • Germany is low. For understandable reasons, World War 2 created strong disincentives against the idea of cultural superiority. This created an aversion to assimilation, which could be seen as a form of cultural domination. Currently, immigrants and resident aliens desire more assimilation than German citizens do!
  • The cultural power of the home country. The higher the cultural power of the home country, the less the immigrant wants to assimilate. For example, some people believe that Chinese students at a 4-year undergraduate degree in a Western university will be filled with admiration for democracy and liberal values. This misunderstanding is based on the assumption that Western countries such as the United States have high assimilation power. What it misses is that China has high cultural power through systematically enforced education, so the students reject the host country’s values.
  • The cultural similarity of the host country and the home country. In 1951, Hong Kong started with a population of 2 million Over the next quarter century, it took in probably 2 million refugees from mainland China. We’d expect a massive involuntary inflow of migrants, equal in size to the entire existing population, to create problems for assimilation. Culturally, though, the immigrants and the host country were almost identical, so the assimilation took place relatively easily.
  • The cultural similarity of the migrant and host country, and the fit between the economic skills of the migrant and the economic needs of the host country.
    • As previously mentioned, these will be high when the level of migration is low, due to selection effect. A good fit leads to high assimilation and positive social capital, because the immigrant wants to fit in with their host country.
    • A better fit with the host country generally means lower fit with home country. So low levels of immigration provide the most marginal benefits to all parties.
    • Voluntary immigration leads to higher self-selection, and therefore better fit. For example, Syrian refugees are a reasonable cultural fit for Europe and have useful skills, but they are significantly harder to assimilate than the Indian middle class going to the United States. “The skills asylum-seekers do have often do not fit local needs. Many newcomers to Germany are experienced car mechanics, for example, but Germany’s shortage is in mechatronics, which requires knowledge of information technology rather than fiddling with fuel pumps.” “But Sweden’s fast-track programmes serve only a fraction of the 110,500 adult arrivals in 2015 and 2016 (around half of whom were granted asylum). In 2017 only a third of refugees who completed a two-year full-time integration programme were working or studying three months later, according to Statistics Sweden. In the first quarter of 2018 this rose to 41%, but only 6% were in unsubsidised jobs. Refugees in Germany have historically fared a bit better. Yet nearly three-quarters of those in work have jobs needing few skills and with poor prospects.”
    • Example: the enterprising husband moving to a country with more liberal markets to the wife he drags along 5 years after he establishes himself.
  • The immigrant’s non-transferable and non-transportable economic and social assets. Significant value is lost in non-transferrable and non-transportable assets, such as local knowledge that is now irrelevant, business and personal relationships that are difficult to maintain, and cultural habits that may be counterproductive in the host country. These only affect the home country and immigrant, not the host country.
  • The amount of immigration into the host country.
    • Higher levels of migration relative to the host country’s existing population leads to more significant degradation of social capital, therefore a higher social cost.
    • Higher levels of migration relative to the home country’s existing population leads to a weaker selection effect, and therefore worse fit, leading to lower economic benefit and higher social cost.
  • Intra-country differences in tolerance of immigration.
    • Even within a country, there is more benefit to immigration for areas with high openness to experience (cities), and less benefit for areas with low openness to experience (rural).
    • Sorting effects mean that many countries have a two-track self-segregated population, with an urban, liberal, high openness part easily assimilating immigrants, and a rural, conservative, low openness part having a tough time.
    • High openness leads to higher benefits from immigration, because immigration increases diversity. Diversity provides benefits for high openness communities, but creates costs for low openness communities.



The base case for immigration is that it provides a benefit to the immigrant, a benefit to the home country, a cost to the host country, and a benefit to the host country that is unevenly distributed. The variation in all of these is significant and is sensitive to the many factors listed above.

Perhaps immigration can be thought of as extreme job retraining. Someone who has worked for decades as a coal miner is unlikely to be able to retrain themselves into being a computer programmer.[iv] Immigration is an even bigger change – instead of training for a new job, the immigrant is training an entire way of life. Experts systematically underestimated the difficulty of job retraining, and now systematically underestimate the difficulty of assimilation. Surely we owe it to ourselves to understand what would maximizes the chances of success before we adopt strong positions on immigration.


[i] The host country can have an effect if it is prepared to adopt draconian measures to prevent immigration. North Korea is an example. China has the state capacity to attempt it. But most countries don’t have the ability to do so.
[ii] The $300,000 figure is calculated the following way: We assume that the last person to stay behind has equal or higher economic benefit to staying than to moving. Then, we estimate the economic benefit of the move for this marginal person, which should be equal to the economic benefit from staying. $300,000 is a rough estimate only, and the marginal migrant almost certainly has a much lower cost.
[iii] This is true in almost all cases; exceptions are when areas of the host country have exceptionally high openness to experience.
[iv] Unlikely, not impossible!