The emphasis seems to be on cultural differences. ERS has an analysis of "nonmetro" counties and population loss. Four of its takeaways: "
- Rural out-migration peaked in the 1950s and 1960s (not shown on graph), but was offset by high "baby boom" birth-rates.
- Net out-migration from nonmetro areas was more severe during the 1980s compared with 2010-16, but overall population change remained positive during the 1980s because natural increase contributed roughly 0.5 percent growth annually (compared with 0.1 percent recently).
- Nonmetro net migration rates peaked during the 'rural rebound' in the mid-1990s and again in 2004-06, just prior to the housing mortgage crisis and economic recession. Net migration remained positive for much of the past two decades, increasing nonmetro population every year but one from 1990 to 2009, but net-outmigration has since contributed to population loss.
- The Great Recession contributed to a downturn in natural increase, as fewer births occur during times of economic uncertainty. But falling birth rates and an aging population have steadily reduced population growth from natural increase in rural counties over time, in line with global trends."
I wonder how much of the cultural differences are due to this sorting? Presumably the people who stay in rural areas are more integrated into the locality, more active in churches and civic organizations, more committed to having a career, or rather, to making a living through local job. While the people who move, who go to college and never come back, those people are more into careers in academia, or finance, less interested in religion, etc.
[Update: the effect of the rural out migration means that existing institutions, the schools, churches, stores, etc. lose vitality and makes it hard to create new organizations to meet new needs.]
A final speculation: note the ERS says that nonmetro areas suffered a net loss of population since 2010--that may be both a symbol and a cause of discontent in such areas, discontent leading to the 2016 election result.