,
15 tweets,
5 min read
Read on Twitter
Two intriguing pieces on measuring & questioning rural America's decline, by @andrewvandam @kdrum, h/t @ModeledBehavior. @andrewvandam kindly shared his formatted data -- I'm grateful. Some thoughts.
washingtonpost.com/business/2019/…
motherjones.com/kevin-drum/201…
washingtonpost.com/business/2019/…
motherjones.com/kevin-drum/201…
These pieces call out a common data error, which makes a big difference in this case. If you compare urban & rural growth (actually metro & non-metro) 1950-2017, it matters a lot whether you use metro definitions from 1950 or 2017.
When comparing growth across groups , use group definitions at the start of the period. To compare metro & non-metro growth 1950-2017, use 1950 definitions. Why? 2017 metro definitions reflect past growth: many fast-growing places got re-classified into metros.
And those pieces are totally correct that metro counties grew slower than non-metro counties 1950-2017, when using 1950 metro definitions. Using 2017 defns shows -- somewhat tautologically -- that places that grew too slowly to become metros grew slowly.
But the faster growth of non-metro counties since 1950 does not mean rural decline is a lie -- nor that rural places are doing just fine, or that this statistical issue is the real reason for apparent rural decline.
Fast growth since 1950 in non-metro counties circa 1950 is driven by newer metros like The Villages FL and the addition of booming outlying counties to Sunbelt metros like Phoenix, Dallas, and Atlanta -- places quite unlike non-metro places today.
A more accurate picture of rural areas today comes from more recent patterns of growth -- of course measured properly with metro definitions at the start, not the end, of the period.
Growth to 2018 from either 2010, 2000, 1990, or 1980 has been faster in metro than non-metro counties -- using metro defns at the start of the period. For instance, cum pop growth 2000-2018:
19% in metro counties
6% in non-metro counties
19% in metro counties
6% in non-metro counties
1950 is too different a starting point. 1950-1980 was a period of rapid national pop growth and urbanization. % living in metro counties (contemporaneously defined) rose from 54% in 1950 to 75% in 1980, and only to 86% by 2018.
On just about every economic measure, non-metro counties today are lagging metro counties. It's not a statistical artifact. Measured properly, non-metro growth has been slower -- for decades. And that's not all.
Unemployment is higher and prime-age employment-population is lower.
Educational attainment is lower in non-metro than in metro counties -- and non-metro counties with less education have especially slower growth.
And -- looking to the future -- industries that typically locate in urban areas are projected to grow faster than typically rural industries. hiringlab.org/2017/10/24/bls…
In short: yes, metro vs non-metro growth comparisons can be done wrong & misleadingly. And, yes, many booming places today were rural in 1950. But: none of that means rural America today is just fine. Urban-rural gaps are a serious economic challenge, not a statistical artifact.
(That was the end. And thank you again to @andrewvandam for sharing the metro-code data and showing how sensitive long-term growth comparisons are to how groups are defined.)
