Taking the Long View on Migration
How Do Modern Americans Compare to their Forebears?
Those that have read my work in the past know my fondness for finding comparison data in order to find new perspectives on current migration. In the past, I’ve discussed how migration has changed in living memory, and some data on migration since 1850. For today’s post, I’m going to take a look at what some old Census population data may tell us about migration. In reality, the data I will use is not migration data: it’s just decennial population data by state. But I’ve shown in the past that differences in interstate migration are the strongest correlate of population growth. For periods where migration data is available, it’s inexcusable to skip the migration data and focus on population. But for periods where such data is not available (like 1790–1850), we have to stick to population data. As I will show, migration has changed radically over the course of American history.
I use Census decennial data that can be found here. My method is as follows: I take the rate of population growth for each state, then take its 10th root to get an annualized growth rate. I then subtract the national average growth rate. The result is a “residual annual growth rate.” This residual would include differences in fertility and mortality, differences in international migration, and internal migration. Again, if I had any other data, I would avoid using this type of calculation: but I don’t. Thankfully, for years where I do have data, my results are pretty close to actual values.
American Migration History
Gross Migration Rates
The above chart tracks 5-year migration rates from 1850 to 2010 as presented in two fairly recent economic papers that I have discussed elsewhere; the authors’ names are listed. I am indebted to their work for this data. Their calculations show fairly high migration in the mid-1800s, lower rates until 1940, then much higher rates in the post-war period (although they have declined some since their peaks in the 1970s).
This chart, which I have also shown before, covers just the last 60 years on a more detailed (and annual) basis. Interstate rates are shown in red, and do indeed seem to have declined in recent years, though we have no data before 1948.
So overall, we can say that American migration is near its lowest point since WWII, but still higher than at any point since before WWII.
American Migration History
Pioneers and Snowbirds?
But wait, there’s something weird about that idea. Native Americans not on reservations were excluded in basically all historical data, so new lands were “empty” for Census purposes. Yet, somehow, they were settled. It had to be migration, right? And what about the Oregon trail, the Wild West, and conestoga wagons? What about the buffalo and fording rivers and getting dysentery (you can tell I played a lot of Oregon Trail as a kid)?
Is the migrant narrative a lie? Was migration to the west actually unspectacular? Are modern migrations to the Sunbelt or the suburbs actually dwarfing those historic migrations?
It turns out to depend on how you look at it. To explain what I mean, we’ll need to brush up on some migration vocab.
American Migration History
Changing Migration Flows Over Time
The above chart shows my estimated “Average Net Migration Rate,” or ANMR, from the 1790s to the 2000s, based on decennial Census population data. This statistic indicates the average impact of migration. Again, for my data, it technically measures the average of all residual effects, ranging from differential natural growth to differential internal or international migration. For this post, as a kind of intellectual exercise, I’ll assume that the whole residual is internal migration: but remember, that’s not actually a true assumption.
In the early 1800s, I estimate that the average state saw net migration of around 4.5% or -4.5%. Some saw far more, others far less. Those peaks you see tend to correlate with successive “frontier population booms.” In the 1810s and 1820s, the boom states are Alabama, Arkansas, Illinois, Indiana, Mississippi, Missouri, and Ohio. In the 1830s, the boom states are Michigan and Wisconsin. In the 1850s, the boom states are California, Iowa, Minnesota, Oregon, Utah, and Washington. And in the 1870s, the booms are in Arizona, Colorado, Idaho, Kansas, Nebraska, Nevada, the Dakotas, and Wyoming.
Frontier states were settled by migration. Although gross migration was low in the 19th century, net migration was high, because migrants moved to low-population areas, and because most migration was one-directional. Nowadays, lots more people move from California to Texas than the reverse, but, even so, there are many people who do move from Texas to California. They’re outweighed by the other direction, but still numerous. In the 19th century, not only were migrants moving into “open land” (of course, Native Americans were there, and violently displaced) but almost nobody was making the opposite flow. When you combine one-directional migration with low population in receiving areas, even low migration rates can create huge migration impacts.
But just how one-directional was migration?
The above chart shows a statistic called the Migration Effectiveness Index for the 1850s to the 2000s. To compute MEI, you have to have both the ANMR and a measure of gross migration. For gross migration, I use the data presented in the first chart, from previous academic studies, adjusted slightly to make it more comparable to 1-year migration rates. By dividing these values, we get a statistic that shows how big net flows are compared to gross flows. A score of 0 indicates that all regions have perfectly balanced migration. A score of 100 indicates that migration is perfectly unidirectional.
MEI for the US peaks in the 1870s at 98.4%: migration was almost perfectly unidirectional. Okay, so the caveats: my data is flawed for such precise measures. I’m definitely picking up some noise from changing international flows, and from the difference between 5-year estimates and derived 1-year estimates. So I wouldn’t actually bet big money on a 98% MEI. In the modern world, it’s extremely rare to have an MEI over 50%, with just a handful of countries exhibiting such lopsided migration.
All the same: it’s almost certain that MEI in the 19th century was quite high, and it’s highly probably that MEI has been falling pretty much ever since.
One way to compare these two trends is to index them both in a common year. I’ve chosen 1850. MEI and ANMR have remained almost identical over the long term. Where they deviate, they do so because of changes in gross migration. While migration was low from the late 1800s until WWI, MEI tended to be relatively high. When migration rates rose in the 1940s and 1950s, similar levels of lopsidedness created a substantially larger net migration impact, so ANMR rose to relatively higher levels.
Migration in the United States has changed drastically over the last two centuries. Although people move more than they used to, those movements are much more balanced, leading to a smaller net impact on regional populations. But while the net impact may be smaller, migration may still effectively reshuffle people based on economic or social factors. A smaller net impact means that migration has less clout in reshaping our population geography, but it doesn’t mean that migration has a smaller role in shaping American lives. An American today is more likely to be a migrant than at almost any point in our past.
See my previous posts, about the migration history and recent trends.
Start the series from the beginning!
A Brief Methodology Note:
Bad data is no substitute for no data. All too often, there is a temptation to force a given dataset to tell a story it simply cannot tell. I readily admit that this post relies on some very obviously incorrect assumptions (for example, that international net migration is evenly distributed around the country), and in that sense it amounts to “bad data,” and thus is no substitute for no data. However, I hope that I made one thing clear through this post: I am not trying to make any kind of “strong” claim here about relative migration rates, just trying to get a general impression of what happened in the past. I’d be prepared to believe ANMR for the period 1790–1880 was anywhere from 2.5% to 8.0% based on my reading of the data. But whether you take the high or low number doesn’t matter that much: because there’s no way ANMR has ever been anywhere close to 2.5% in the 20th century. It would take some genuinely incredible and even heroic assumptions to arrive at at such a number. And the ANMR number I arrive at with my method is actually pretty close to Census Population Estimate numbers for 1990–2014. Likewise, do I think MEI was 98.4%? Certainly not. Was it almost certainly above 40%, probably over 50%? I think that’s hard to dispute. And since 1920, it has not approached those levels again, not even close.
So my point is this: bad data won’t substitute for no data in showing what’s true. But sometimes even bad data can help you rule out claims that are definitely false. And the claim, for example, that “US migration has had about the same level of ANMR or MEI from 1790 to present” seems almost certainly false.
If you like this post and want to see more research like it, I’d love for you to share it on Twitter or Facebook. Or, just as valuable for me, you can click the recommend button at the bottom of the page. Thanks!
Follow me on Twitter to keep up with what I’m writing and reading. Follow my Medium Collection at In a State of Migration if you want updates when I write new posts. And if you’re writing about migration too, feel free to submit a post to the collection!
I’m a graduate of the George Washington University’s Elliott School with an MA in International Trade and Investment Policy, and an economist at USDA’s Foreign Agricultural Service. I like to learn about migration, the cotton industry, airplanes, trade policy, space, Africa, and faith.
My posts are not endorsed by and do not in any way represent the opinions of the United States government or any branch, department, agency, or division of it. My writing represents exclusively my own opinions. I did not receive any financial support or remuneration from any party for this research. More’s the pity.
