What Percentage of Immigrants Are Actually Immigrants?
Less Than You Might Think
Longstanding readers know I’m very interested in estimating good data for actual migration flows, which is harder than you might think. And sometimes it’s harder than I might think. This post is a revision to some of my survey-based estimates, reflecting the fact that I did not adjust for the U.S. citizen share of inflows, because I woefully underestimated how big a share of reported inflows that might be.
One quirk of the immigration data we have that interested me was that we have inflow data broken out by citizenship status for at least some years, going back to 1940. We have ACS data back to 2001, CPS data back to 1994, and we have decennial census estimates in 1950 on a 1-year migration basis, and for 1940, 1980, 1990, and 200 for a 5-year migration basis. Conveniently, for estimating the share of the immigrant flow that is of a certain class, the 1-year/5-year migration time window shouldn’t matter very much. Here’s each data source, with linear extrapolations made for missing years.
What do we notice here?
First of all, U.S. citizens used to be a way bigger share of inflows than they are today. Second of all, we’re returning to that norm! The U.S. citizen share of immigration is rising! Today, at least 1/3 of all inflows from abroad into the U.S. in the last year are U.S. citizens.
In my rundown on immigration statistics, I provided a “survey based” estimate of immigration. However, it included U.S. citizens as immigrants! That’s a mea culpa! So here’s that data, shown again, but reducing immigration estimates based on the share of “immigrants” who are actually U.S. citizens. I take the average value of my three indicators for years where I have multiple indicators, and add smoothing for the introduction of new indicators.
As you can see, the noncitizen immigration rate is much lower. This changes some of the conclusions from my immigration factsheet post, but not all. For example, peak post-war immigration shifts to being in 2001 instead of 1970. I compared survey-based immigration to category-based immigration based on green card offerings and an estimate of illegal immigration in a chart. Here is that chart, updated:
As you can see, removing U.S. citizens from the calculation on inflows dramatically reduces annual immigration rates below the levels estimated in either the raw survey or category-based methods. This changes a key conclusion. I said in the old post:
These numbers are not quite right. Average inflow rate for post-1940 remains 0.51% for the survey method and 0.57% for the category method, but for the adjusted survey method, the average inflow rate is 0.28%, a much lower long-run migration rate. Meanwhile, adjusting for U.S. citizens creates a trend that is more similar to the category-based method (for the nerds, the raw survey-based measure has an R2 vs. category based of ~13%, while the adjusted measure has an R2 of ~50%).
Using the new survey measure, current migration rates are about 26% below their peak, which is not quite as low as I originally estimated, but very similar.
There is a problem, however. This new estimate yields migration rates that are hard to square with category data that we “know” occurred. That is, we know that X green cards were issued. Here’s annual green card issuance versus my adjusted survey-based immigration:
As you can see, my adjusted measure is, in the long-run, substantially above green card issuance, so it passes a basic sanity check. For reference, total immigration 1940–2016 would come to about 52 million vs. total green card issuance of about 44.5 million, meaning non-permanent or illegal entrances of about 7.5 million. For reference, my estimate from border apprehensions is more like….61 million people. Yes, I estimate we’ve had more illegal entries than legal ones. But I also estimate that emigration of illegal entries is much higher than of legal entries. There’s also some weirdness with the Bracero program, but that’s neither here nor there.
The point is, either the major surveys of migration are missing a lot more illegal entrants than they ought to be (whether that’s due to missing Hispanic population, or false responses regarding timing of entry by illegal entrants, would be an open question), or the relationship between illegal entries and apprehensions is not as clear cut as many observers treat it. I suspect some of both is happening. Probably the surveys miss some people and probably my 61 million number is too high. I do suspect that CPS/ACS/Census probably systematically miss “high-churn” migrant groups like seasonal workers or workers intending to stay only a few years. Apprehension data would capture those migrants better. So this is kind of a subjective question: should we count people who do not intend to be permanent, and who maybe 75% of the time leave within 1 or 2 years, as part of the “population” that CPS/ACS/Census track? Is this an important group to know about? Should we count their inflows/outflows in headline “immigration” data?
But regardless, relative to my previous post, this should cause you to (1) revise your estimate of total migration downward slightly and (2) revise your estimate of when we hit peak migration levels forward in time slightly.
Check out my Podcast about the history of American migration.
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 native of Wilmore, Kentucky, a graduate of Transylvania University, and also the George Washington University’s Elliott School. My real job is as an economist at USDA’s Foreign Agricultural Service, where I analyze and forecast cotton market conditions. I’m married to a kickass Kentucky woman named Ruth.
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.
