My son Daniel points me to a piece by Nate Silver which announces the discovery that highly educated workers with low or precarious income are the natural constituency of the Democratic Socialists of America. We’ve been over this at length at CT, starting as far back as 2012 (and it wasn’t new then). Given that Silver treats the question as new, I thought it might be useful to restate the basic points from this discussion.

• Always and everywhere, controlling for education and other variables, higher income is correlated with rightwing voting
• Nearly always and everywhere, controlling for income and other variables, more education is correlated with leftwing voting
• Because income and education are positively correlated, these two tend to cancel out
• With the (now only historical) exception of parts of the US, rural voters in developed countries are reliably rightwing
• Because of the history of slavery, civil war and segregation, US, Southern Whites in the US vote like an ethnic bloc, defined in opposition to Blacks. This fact largely accounts for the greater success of Trumpism compared to European far-right parties
• The most important form of identity politics is default identity politics, based on the idea that members of the unmarked category (in the US, white Christian men) are the natural representatives of the country and should therefore run it
• All of these are tendencies rather than uniform rules. There have always been “working class Tories”, “millionaire liberals” and so on

A nearly final observation following from this. As a conceptual tool for analysing voting patterns, the term “working class” is too ambiguous to be useful. It can mean anything from “wage workers in labouring and manual trade occupations” (perhaps 20 per cent of the population) to “below median income” (50 per cent by definition) to “did not complete 4 years of college” (about 60 per cent) to “income not primarily derived from capital ownership” (95 per cent), with only partial overlap between these categories. All of these definitions are useful in some contexts, but in debates over voting patterns they form a hopeless jumble.

And finally a restatement, While I’ve linked all of this to my own posts, I’m not claiming originality for these ideas. Just making the point that Silver is claiming as a discovery something that anyone paying attention could have observed at least a decade ago.