The 1920s and H1B visas

A question scholars of historical political economy frequently get from those outside the field is why we need historical studies to answer questions that could also be addressed in modern contexts. One is justified to ask this, as many of us are not historians, but social scientists employing history to draw generalizable conclusions about human behavior and institutions. If we care about solutions to today’s burning problems, why not focus on those problems, rather than looking for answers a few decades or centuries back? Why prefer sparser, noisier (and endogenous) historical data over often much larger, more detailed and error-free modern datasets?

Two answers frequently come up (though there are more than two). The first one is that history uniquely allows us to study phenomena that manifest over the (very) long-run, such as intergenerational mobility across several generations, or the long-term effects of institutions. The second one is that history is a particularly useful place to look when contemplating decisions of a public policy nature.

Pandemic management is a good example of this relevance of history for public policy, as Ali Cirone explained in her recent post. Another example is immigration. In the last couple decades immigration has transformed the political landscape in the US and Europe. It tops voters’ lists of policy priorities, and it features as a major point on every party’s agenda – particularly the agendas of increasingly popular right-wing nationalist parties, like the German AfD, or the French Front National.  Debate is especially heated around policies to restrict immigration (like Donald Trump’s border wall with Mexico), as well as policies to deal with diversity once immigrants are already a sizable minority in a country (like ghetto laws in Denmark).

Turns out that surprisingly similar debates have taken place in the past, often in the same countries that experience them today. Their outcomes, and the decisions of past policy-makers conveniently provide us with a menu of counterfactual realities to draw from, to inform our modern-day policy decisions.

Take for example immigration restriction. At the beginning of the 20th century, after receiving nearly 30 million European immigrants and experiencing decades-long intense debates around their impact on the country’s economy and society, the United States decided to halt their entry. The 1917 literacy test, and the 1921 and 1924 national origins quotas were designed to severely restrict the numbers of immigrants from Southern and Eastern Europe, and exclude those from Asia. With the quota system, the US transformed from a country of open borders and a home for the world’s tired, poor and huddled masses, to a destination closed to all but a small number of white Northern and Western Europeans.

A recent NBER working paper by Ran Abramitzky and coauthors examines the causal impact of the 1920s quotas on the American economy. The impact of the quotas was most pronounced in places with large numbers of Southern and Eastern European immigrants. Migrants tend to move to where their families and co-ethnics live, and so the immigrant population in areas with large enclaves of Southern and Eastern Europeans would have grown much more were it not for the restrictions. Comparing those areas before and after the introduction of the quotas, the authors document a decline in the income of US-born workers.

This finding goes against the prediction of a simple neoclassical economics model that a decrease in labor supply should increase wages. Instead, it accords with a slightly more sophisticated mechanism, in which firms respond to the drop in supply by shifting to other forms of labor and capital. In the case of US manufacture, the loss of foreign workers was entirely compensated by in-migration of local residents and immigrants not affected by the quotas (Mexican and Canadians). US farms shifted instead to more capital-intensive forms of production.

This study highlights two things. First, the effects of immigration restrictions are heterogeneous. Some domestic workers benefit, for instance by taking the jobs immigrants leave behind. In related work, William Collins shows that, had it not been for immigration from Europe, Black Americans would have moved out of the US South, and to economic opportunity in the North, several decades earlier than they actually did.

Second, the aggregate effects of restrictions on the economy are either null or negative. This is confirmed by Tabellini (2020), who shows that areas less impacted by the quotas saw better economic outcomes, as measured by incomes and value added in manufacturing. In a recent working paper, Moser and San document the negative impact of the quotas on US science and invention.  Clemens, Lewis and Postel (2018) examine a different type of immigration restriction, enacted in the 1960s: the end of the Bracero program that brought Mexican agricultural laborers to the US for seasonal work. Like Abramitzky et al., they too find that the program led American farmers to mechanize their production methods, and had no effect on the wages or employment of domestic agricultural workers.

Based on history then, curtailing immigration does not seem like a great idea, at least in terms of its aggregate implications. Historical work provides one additional reason to rethink immigration restrictions: the effects of immigration in the long-run appear to be overwhelmingly positive. Sequeira, Nunn and Qian (2020) quantify the exposure of US counties to immigration between 1850 and 1920. Counties that were connected to the expanding railroad system in decades that saw higher outflows of immigrants from Europe were more likely to receive a larger number of immigrants. Those counties experienced greater industrialization and higher productivity that affected their growth trajectory for up to 100 years. In 2000 the population of those counties was richer, more urban, and more educated. This did not come at the expense of lower social capital, or higher crime rates, negative outcomes often associated with ethnic diversity.

Despite the predominantly positive effects of immigration, backlash against it has been as intense in the historical past as it is today. One might think that was primarily motivated by the short-run negative economic effects of immigration on segments of the native population, and there is clearly some truth to that (see, for instance, this recent working paper by Alsan, Eriksson and Niemesh on the success of the Know-Nothing party). Yet historically, like today, cultural and social considerations were equally, if not more important. Tabellini (2020) shows that, while Catholic and Protestant immigrants had very similar – and largely positive – effects on the economy of US cities, only the former provoked intense political reactions. Catholic and Jewish immigration led to curtailing of redistributive policies, election of more conservative legislators, and more support for immigration restriction. This finding is not context-specific – much work in political science today confirms that, when it comes to attitudes towards immigrants, cultural factors trump economic ones. Native populations are averse to newcomers, especially when those are culturally and religiously different.

So, what lessons can we draw for today’s public policy debates? Clearly, the answer is complicated by the fact that policy agendas on immigration are mainly dictated by the power of different groups and politicians’ incentives for re-election. But all the aforementioned studies provide a hypothetical social planner with crucial information for policy design.

One insight that emerges is that policies targeted at mitigating negative reactions to social and cultural diversity can be fruitful. Research suggests that sustained interactions between different groups in positive contexts can lower hostility, which points to the importance of housing policies and residential integration. Perhaps the clearest lesson is that immigration, like trade, grows the aggregate pie. Public policy debate may need to move away from immigration restriction to protect group-specific interests to assessing redistributive strategies that ensure everyone participates in the aggregate gains.


Leave a Reply