Who’s Migrating to the Sunbelt?

It’s well-known that people have been moving from coastal US states to the Sunbelt for many years now. But who’s moving? Is it the upper middle class fleeing higher taxes or searching for cheaper houses, or perhaps the poor fleeing high costs of living? Put another way, is the above-average growth in per capita income in many Northeastern and West Coast metropolitan areas a matter of actual growth, or simply of pushing the poor out to the Sunbelt, whose per-capita income growth is often anemic?

All data in this post is courtesy of Aaron Renn’s Telestrian service, which cribs numbers from the IRS, Census Bureau, and other sources and presents them in a reasonably searchable manner. The IRS keeps track of intranational migration in the form of tax exemptions, which allows us to figure out the migration trends in terms of people (exemptions), households (returns), and income (adjusted gross income). This way we can figure out if the people moving out of a region are richer or poorer than the average. Although the IRS misses a lot of people and much income, it is still the best available source in the US for migration statistics. The more accurate American Community Survey tabulates very coarse migration statistics.

Observe also that the IRS data is given per year, which allows us to look at zoomed-in trends. For example, here is California’s migration with each state as well as the rest of the world from 2000 to 2009. Here is somewhat worse-presented data for New York State. It turns out that migration marginally increased California’s per capita income, and had practically no effect on New York’s; in other words, their growth is real, and doesn’t come from pushing the poor away.

More precisely, we have the following observations:

– In both California and New York, the difference in income between immigrants and emigrants is very small; immigrants are slightly richer in California, $27,098 vs. $26,209, and slightly poorer in New York, $29,876 vs. $30,810.

– Overall both immigrants and emigrants are slightly poorer than the statewide per capita income. However, the effect is very small: according to the IRS, California’s per capita income in 2009 was $28,569 and New York’s was $31,617. Were it not for migration – that is, if people had lived in 2009 where they’d lived in 2000 but still earned the same income – California’s per capita income would’ve been $28,243, i.e. 1% lower, and New York’s would’ve been $31,689, i.e. 0.1% higher.

– The richest migration occurs between high-cost coastal states, especially between New York and California, while migration between those states and the Sunbelt is much poorer.

– The poorest large group of immigrants in both states is international immigrants. In both cases they were about 9% of total immigrants, so they can’t have dominated the numbers too much. Thus Jane Jacobs’ story that great cities take in poor immigrants and churn out a middle class, considered on the state level, is only partially confirmed by this data.

– Emigration to the Sunbelt’s bubble states – from California to Arizona and Nevada and from New York to Florida – was predominantly a 2005-7 phenomenon, and decreased markedly after the bubble crashed.

– Emigration to other Sunbelt states was more of a mixed bag. Georgia and North Carolina, both partial bubble states, also look like partial bubble states in the migration numbers, with emigration from New York and California peaking in 2005-8, but less prominently than with the proper bubble states. Emigration from California to Texas looks like that to a bubble state, despite Texas’s strong economy through the recession; but emigration from New York to Texas and from both states to Colorado remains steady.

– The biggest difference between immigrants and emigrants is not income but household size – emigrating households were much bigger (1.95 vs. 1.7 in California), but still much smaller than the statewide average (2.23 according to the IRS, much lower than the actual average but comparable with the above numbers). This is only partially consistent with the explanation that those regions attract singles and DINKs and turn away families.

The story I started this investigation with is that New York and California predominantly turn away the middle class, which would be seen in middle-class emigration and low-income immigration; my recollection, coming from merely eyeballing the data, had been that immigrants were much poorer. This should be consistent with the breakdown of the cost of living in dense city regions: housing is unaffordable if your ideal of how to live is having a car and a single-family detached house that’s less than an hour away from work; if you’re flexible about car ownership and don’t mind small apartments, then New York and California are quite affordable.

But what we actually see is that both immigration and emigration between those states and the rest of the world is middle-class. The people moving to the Sunbelt really are being priced out. It’s hard to distinguish pricing out from cashing in on high housing prices, but the lower-income characteristic of this emigration suggests the former. The upshot is that policies reducing the cost of housing could stem this tide while at the same time having no effect on poverty and the need for social services. While it’s heinous to try to price out the poor, as the richer parts of the Bay Area and many other regions do, this is not what is being done here.

Let me close by linking without much comment to the same data for Texas. The IRS recorded a total income of $475,109,477 in 2009 and a total population of 19,235,926, i.e. a per capita income of $24,699. As in California, immigrants are a little poorer than emigrants and both are a little poorer than the average. Controlling for this effect as above would raise per capita income to $25,002, a 1.2% rise.


  1. Andre Lot

    Housing policies, alone, cannot hold population in place, Alon. Just look at many Midwestern cities that, despite extremely cheap housing prices, have been loosing population since the 1970s (at the latest). One must follow the jobs trail as well: construction booms attracts construction workers who are generally poorer then the pop. average.

    As for the specific case of housing, I’d not compare the situation of NYC core metro area with anywhere else. Living car-less in NYC might be possible, and demand less compromises in terms of mobility and being effectively shut out many opportunities than living in any other metro area, including Bay Area, Los Angeles etc.

    • Alon Levy

      Of course they can’t. Not all Rust Belt areas (counting coastal California as Rust Belt) are created equal. The big core cities are desirable; the smaller metro areas aren’t. But in the big cities, the high demand for housing has translated entirely to higher rents and property values rather than to high population growth.

      I didn’t think of the explanation about construction workers specifically. It could be right, at least for bubble states. On the other hand, if we look at the income growth of the bubble metro areas, in some it was not much lower than the national average, e.g. Orlando and Tampa.

  2. mlewyn

    I’m not sure how useful statewide data is. New York includes coastal, high-cost NYC, but also includes low-cost Central and Western New York. California includes the high-cost Coast and less expensive inland areas. Now, if we had similar data for counties, then we might learn something!

    • Alon Levy

      True. But, county-level data is not that useful either, because it’s contaminated with city-suburb migration. This is partly true of NY-NJ and NY-CT moves, too, but not as much as for individual counties (and of course, it’s not an issue with California).

      The reason I wanted to include Upstate with this bundle is that the fastest per capita income growth was Upstate. New York’s metro area only grew about as fast as the national average; the three major Upstate metros (omitting Albany) grew about twice as fast.

      There’s MSA-MSA data I could look at, but to be honest I doubt I’ll see anything much different. California’s population is concentrated in the big coastal metros, counting the Inland Empire as part of Greater LA, and a significant chunk of the rest is in Sacramento, which doesn’t count for much any more than Albany does.

  3. Dave Roberts

    It’s amazing how many planning and zoning personnel don’t understand supply and demand. They literally can’t make the mental leap from down-zoning a parcel for fewer homes to an increased cost of housing. My favorite example is BART stations in the Bay Area which are surrounded by detached single family homes rather than high density dwellings. Governments have to realize how much our land use planning has impeded the highest and best use of property. I endorse smart planning, but what we have today is often the worst of all worlds.
    With smart changes to planning policies, we could fit millions of additional people into the greater Bay Area without consuming one additional acre of land for new construction. That’s how you make things affordable.

    • Waingro

      In fairness, it’s not really professional planners per se that are guilty of supply-reducing policies. It’s generally incumbent homeowners trying to pull up the ladder so outsiders/riff-raff can’t lower their home values. These residents then elect the zoning boards who pass supply restricting policy. Since planners are generally civil service staff, they’re not really in a position to overrule the elected body.

      Otherwise, I’m in complete agreement. It’s infuriating to see low-rise development next to transit stations.

  4. Pingback: California should fear high housing prices | Lala24 Hot News from all popular web sites
  5. Pingback: Materias Grises » Blog Archive » Si la vivienda es cara, probablemente no hay suficientes viviendas
  6. Pingback: Zoning out the middle class in California because of high home prices – The psychological and demographic impacts of an unaffordable housing market. » Dr. Housing Bubble Blog
  7. Vinnie

    A thought…

    Would this data lead to different interpretations if controlled by the standard of living for the area? The fact that immigrants and emigrants have a roughly similar income (whether by household or per-capita) shouldn’t be surprising on a macro level, especially if using IRS data — usually when someone moves, they aren’t taking their job with them, so a newcomer gets the same / a similar job. However, if there was a way to tally incomes of immigrants/emigrants based on their destination income and cost of living (perhaps relative to a locally-variable poverty line, basket of goods, or other such benchmark), you might see the picture that is the more familiar stereotype.

    That is, that emigrants from high-cost states are moving to situations where they can end up living middle class, while the immigrants replacing them are moving to situations where they qualify as what researchers now call “near-poor” (what we used to call “working class” or basically paycheck-to-paycheck) or worse.

    Completely agree with you that policies reducing housing costs are the solution, though.

  8. Pingback: California should fear high housing prices « Sunshine Properties's Blog

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.