How Big is Migration?
Bigger Than in Most Countries, Smaller Than in the Past
Migration is a powerful force in American society. As I showed in my previous post, interstate migration is larger than international migration in the United States, and by a wide margin. But that doesn’t really answer the question of how big (and implicitly important) migration is: what’s the right benchmark?
There are basically two options for comparison: we can compare to the United States’ own history, or to other countries. I’ll begin with American historical migration.
The Current Population Survey, the most widely-used and long-standing regular survey that asks about annual migration, goes back to the 1940s. After correcting for some errors in collection in recent years, economists Raven Molloy, Christopher Smith, and Abigail Wozniak present the following data in their paper exploring declining migration:
From that data, it seems pretty conclusive that migration is declining, and fast. Interstate migration has fallen by half in 20 years, as has intercounty migration. Intracounty migration has fallen slightly less, but shows the same general trend.
By this measure, we could say that migration in the U.S. is near its lowest point since WWII.
But are those peaks measured really the best comparison? What does the history of American migration look like in the very long term? We don’t have the same survey data, but the Census has asked questions about migration since the 1800s. The most comparable data is 5-year migration trends. Two different NBER working papers, one by the same authors noted earlier, and the other by economists Joshua Rosenbloom and William Sundstrom, provide good data on long-run historical migration trends.
Rosenbloom and Sundstrom impute trends based on childhood birth and residency status, and so are able to build a dataset stretching back to 1850. Thus, their results allow us to look at migration rates even during the settlement of the west and the waves of international immigrants in the late 1800s.
As can be seen, while migration rates have leveled off somewhat in recent decades, they remain higher than at any point before the Second World War. We are a nation of interstate migrants more than we were during the period of the Oregon Trail, the Homestead Act, and the Pony Express.
While migration is near a low point in living memory, it remains far higher than it has been for most of the history of our nation.
But is the United States special in this regard? We can’t really claim any special migration narrative as a nation if our migration rates are actually low compared to other countries. Getting consistent data on internal migration around the world is tricky, but Molloy, Smith, and Wozniak have done the work again here.
This data tracks all residency changes, not just interstate movers, so it doesn’t strictly capture what we might consider to be “migration.” All the same, it is suggestive of a trend: Americans move more than almost any other developed nation, and those that have higher residency change rates are very small, culturally homogenous Nordic countries that don’t have as many possible long-distance moves.
If we want a broader range of countries and a definition of “migration” more like “interstate migration,” getting data becomes even more of a chore. One notable project attempting to do this is called the Comparing Internal Migration Around the Globe project (IMAGE). While the IMAGE data is somewhat more dated, as it is restricted to major decennial census data in most cases, it does offer a wide cross-section of many countries, developed and developing. A selection of those countries is presented below.
When comparing similar state- or province-level internal migration in other countries, the US’ internal migration rates appear extremely high, whether 5-year or lifetime migration is measured.
The other two variables presented above, Annual Net Migration Rates and the Migration Effectiveness Index, measure how much of an effect migration actually has on redistributing total population; i.e. do many regions have major negative migration, and others have major positive migration. ANMR measures this effect as a share of the total population, showing how much of the population is redistributed, on net, in the 5-year period, while MEI measures how lopsided migration flows are, regardless of how big migration is compared to the population. On ANMR, the US is in the middle, on MEI, the US is near the bottom. This suggests that the US’ migration rates are roughly balanced, with most outmigrants from a given state replaced by new in-migrants, unlike in China, for example, where migration is almost unidirectional from rural inland provinces to urban areas in the east.
Other specific datapoints support the case that American migration is exceptional. Just 3.2 percent of all EU residents were born in an EU state other than their current residence, as of 2010, suggesting a migration rate possibly as low as 1/10 that of the United States. In Mexico in the same year, about 80 percent of the population resided in their state of birth, suggesting an internal migration rate about 1/2 or 2/3 that of the United States. Across our northern border, in Canada, annual migration rates just recently reached 1 percent, a historic high, yet lower than at any point in modern U.S. history.
“Few countries can match the American propensity to move around within our own borders.”
In China, interprovincial migration is enormous, despite major legal barriers erected through the hukou system. Major “sending” provinces such as Chongqing, Guizhou, Henan, and Sichuan have suffered net out-migration equal to 10 percent or more of their “registered” population as of 2011, while in “receiving” provinces and regions like Beijing, Shanghai, Tianjin, Guangdong, and Zhejiang, up to 30 or 40 percent of the local population can be “registered” to another province. These statistics may have only limited reliability, like much data coming from China, but are suggestive of the large (and mostly rural-to-urban) migration within China. Yet even so, the seismic changes in China due to economic growth and urbanization have led to the most migrant-dependent places having total internal migrant stocks about equivalent to the average across the United States.
I could offer more examples, and undoubtedly some country somewhere has higher internal migration. Some countries also have high “international” migration that reflects customs unions, or previous legal integration, such as Russia, Belarus, and Kazakhstan.
But few countries or regions can match the American propensity to move around (and often move long distances) within our own borders.
So how big is migration in the United States? That depends on your perspective. If we limit the frame of reference to the post-war period, then migration seems very low. But when we look throughout American history, it turns out that these “low” rates are still higher than in the days when Americans were first building our national narrative of migration by going west and settling the plains and mountains. And when you compare international trends, American internal migration appears genuinely exceptional, in no small part because there are very few countries as large and as populous as the U.S. with so few legal, linguistic, geographic, or cultural barriers. But I’ll explore reasons for migration later. For now, it’s sufficient to say that migration is too big to ignore.
Go to the next post in the series!
Start the series from the beginning!
- I am deeply indebted to the authors of the papers I link, and encourage interested readers to read those papers for themselves. Likewise, if you use the charts I provide for anything particularly analytic or publishable, please contact and cite the folks who did the hard work, not me.
Follow me on Twitter. Follow my Medium Collection at In a State of Migration. I’m a grad student in International Trade and Investment Policy at the George Washington University’s Elliott School. I like to write and tweet about migration, airplanes, trade, space, and other new and interesting research. Cover photo from Unsplash.
