Falling Migration Raises Economic Mobility Concerns
Thoughts on Scott Winship’s Recent Paper
I got into a Twitter discussion with Scott Winship of the Manhattan Institute today regarding his recent paper, “WHEN MOVING MATTERS Residential and Economic Mobility Trends in America, 1880–2010”. I mean seriously, with that title, how could I not bite? Regular readers will recall that I’ve made no bones about my feelings on migration being pro-mobility. Overall, Scott’s paper was excellent, confirming what many of us already believed: migration is strongly associated witheconomic mobility. People who move across state lines during early adulthood were overwhelmingly more likely to have higher income or earnings. There are lots of nuances and caveats, and Scott does not provide, and emphatically does not claim to provide causal evidence of a migration-mobility link. But when you look at his charts, it’s pretty hard not to think such a link exists somewhere. I won’t steal his thunder, so go read the paper yourself for more info.
But I had a few problems with his approach. Namely, he offers this remarkable set of claims and implications:
In recent years, observers across America’s political spectrum have expressed concern over declining residential mobility and its implications for economic mobility in the United States… The reality is more complicated… While the share of Americans having moved in the previous year has fallen since the 1970s, this paper finds that other types of residential mobility are now as high as they have been in 100 years or more… this [pro-mobility] connection is much stronger when focusing on the kinds of residential mobility that have not declined — moves between birth and adulthood (and most likely between adolescence and adulthood) and moves across state boundaries. (Emphasis added)
Anyone who’s read my blog can see where I came to a screeching halt.
There are kinds of residential mobility that have not declined? Really? News to me. What kind?
Oh, right, interstate migration.
Wait, what? Nope. Not so:
So interstate migration has definitely declined. Remember, that IRS jump at the tail end there doesn’t reflect an increase in actual migration, but a huge methodology change.
How can Scott have missed this obvious issue? He’s a careful and meticulous researcher, as anybody who regularly reads his work (ooo, me, pick me!) knows. The devil here is in the details. Scott says it’s about the migration that matters.
What Migration Matters?
Scott claims that the most important migration for future life outcomes is migration between adolescence and 30-something. He presents lots of data to that effect, and his case is reasonably convincing. It doesn’t match what Raj Chetty et al. have found about childhood migration, which suggests birth-to-adolescence is very important, but oh well, that doesn’t worry me for now.
When Scott says “migration hasn’t declined” he’s referring to a few key statistics. I’ll offer here three charts he provides that I think illustrate his claim fairly nicely:
From left, these charts show:
1 As of 2000, the percentage of 30-somethings who’d moved to a different state over the previous 5 years was about the same as in 1990, a bit lower than 1970 or 1980.
2 That the share of all adults who’ve moved between states since adolescence has risen from 23% to 24% from 1992 to 2010.
3 The share of 30-somethings who’ve reside in a different state than they were born has declined only-very-slightly since peaking in 1980.
These are key stylized facts Scott relies on to say that migration that really matters hasn’t declined.
But are those statistics meaningful? That’s the core of my debate with Scott.
It’s the Life Cycle, Stupid
(Scott’s Not Stupid, It’s Just A Saying From Almost Before I Was Born: Take That You Geezers)
I told Scott on Twitter I found his claim of stable migration “lol-worthy.” Okay, a bit flippant, I’ll admit: sorry, Scott. It’s Twitter. But honestly, the issue to me seems like a fairly clear-cut case.
See, migration has a life cycle. You don’t have the same likelihood of moving when you’re 60 as when you’re 16 or 26. There’s a pattern. I’ve talked about it before here, here, and you can see it visualized in brightly-colored bars here.
So migration peaks for people who are 18 to 24 years old. That’s also the transition period Scott identifies as the period when migration as the strongest impact on future life outcomes. To be clear, we can join these stylized facts together and say that the period from high school to age 30 has both the highest likelihood of migration and the largest impact of migration on future life outcomes.
Let’s return to Scott’s claims. He measures the likelihood that 30-somethings live outside state of birth in 2010. Well, if you’re 35 in 2010, you’re past your migration-prime, which was 18 to 24. You would have been 18 to 24 between the years of 1998 and 2004, right?
So for the 30-somethings Scott measures in 2010, a disproportionate share of their accumulated stock of lifetime migrations will have occurred between 1989 and 2004. I got those numbers by looking at when a Year-2010–30-Year-Old would have been 24 (2004), and when a Year-2010–39-Year-Old would have been 18 (1989). So then the “prime migration years” for Scott’s 2010 cohort are 1989 to 2004. Practically speaking, then, the prime migration years for these people are the 1990s. Well, what was migration for 18 to 24 year olds like in the 1990s?
I don’t have complete data available at hand, and I want to portray my thoughts rather as I had them while reading Scott’s paper, so here’s the comparison I had in my head:
When 2010’s 30-somethings were at peak migration, their migration rates were way higher than 2014’s peak-migration-age individuals. This should tell us something.
When Scott measures the share of a given age group who have ever migrated during a given window, he’s measuring a stock. But when policymakers say migration is low, they’re referring to a flow.
To be clear, when Scott measures the post-adolescence migration for all adults, that’s an even more lagged indicator, an even stockier-stock. Migration rates were way higher when some of the now middle-aged people where at the point in the life cycle where migration is most likely.
What Year Do We Care About?
So when Scott says that migration hasn’t fallen, he’s actually measuring a 2010 stock of migrants that measures acts of migration that disproportionately occurred between 1989 and 2004. Bear with me people, I know this is painfully complicated, and I promise there’s a reward at the end. He then assigns that value to 2010, and says, voila, migration has not fallen (much —if you look closely you’ll note there actually has been a birth-to-30s decline from 1980 to 2010).
But to quote Obi-Wan Kenobi, that’s only true “from a certain point of view.” Migration had not fallen (much) by the period in question. Look back at the CPS data in the first chart: interstate migration fell only slightly from the 1970s to 1990, and huge amounts of the total decline have occurred since 2000, or even since 2004. By weight, the most significant years for 2010–30-somethings are likely to be in the mid-1990s, when migration rates were almost double their current values.
I completely agree with Scott that migration hadn’t fallen by very much as of the late 1990s. And that’s basically what his data reflects.
We Care About the Future
But when we talk about declining migration, we’re worried about the future. The fact that peak-age-migrants have half the migration rate they had in the 1990s obviously doesn’t show up for post-peak migrants in 2010. It’ll show up in post-peak migrants in 2020, or 2030. The most deleterious effects of a lower-migration society have not yet begun to be felt. They’ll be felt in 10 years when there’s a whole generation of college graduates who lived at home for years, didn’t work, and took their 4th college choice to stay closer to home instead of the program that would have challenged them a bit more.
In sum, Scott’s data isn’t wrong, it just doesn’t say what he thinks it says about the state of migration. It bears emphasizing: This has no impact at all on his findings concerning mobility. This is a tertiary point wholly separable from the rest of the paper.
If you read Scott’s paper and what you get from it is that migration is reasonably likely to promote economic mobility, that higher levels of migration are likely to be desirable, and that there exist a wide gamut of policies that could potentially boost migration, then my critique has no impact on how you read the paper. All of those readings remain intact. However, if you read Scott’s paper and what you get from it is that migration has not fallen, and so there’s nothing to worry about, then you absolutely need to think again. As should Scott. The vast majority of declines in interstate migration have occurred outside the window of time that most drives his results, because his results essentially measure stocks. Lower rates of migration will take time, decades probably, to fully show up in stocks.
And if migration rates bounce back up, currently lower rates may never show up in stocks. Peak-age migrants during the low years remain young enough a change in conditions could induce life cycle catch-up of sorts. Scott asked me if I really think we’ll see birth-to-adolescence rates similar to what we saw in the 1940s or 1950s. Frankly, I don’t know, but if annual rates remain where they currently are then I think it’s inevitable we will eventually see 1940s and 1950s level of migrant stocks. The only way we could maintain currently low rates without lower stocks is if the whole reduction in rates comes from reduced repeat-migration. I don’t think either Scott or myself has information off hand about the average lifetime number of migrations for the whole population or for specific age cohorts, so we’ll both probably have to confess ignorance on that vital statistic for the time being. But, given that limitation, I think it’s fair to say persistent low annual rates will eventually lead to migrant stocks at low levels similar to those seen in the immediate post-war decades.
See my previous post, about which metro areas saw statistically significant migration changes in 2014.
Start my series on migration from the beginning.
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. 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. More’s the pity.
