There's a lot of news coverage today of new research from Raj Chetty and Emmanuel Saez indicating that social mobility in the United States is not falling, offering the not-so-reassuring news that the reason it isn't falling is that it's been low for a long time.* At least 50 years.
Recall the old old conventional wisdom on this was that the United States might be a society of high income inequality, but at least it had a lot of mobility. Then the new old conventional wisdom became that this was wrong—that in international comparisons more egalitarian countries (the Nordics) had more mobility, so as America has become more unequal we've also become less mobile. Now today we get the new conventional wisdom, which says that America is a low-mobility country but has been this way for a while.
I'd like to put on the table a different research program, associated with UC–Davis economic historian Gregory Clark, which argues that economic mobility is low almost everywhere.
He reaches this conclusion with a different research method that lets him explore much longer-term trends than most of the research you see on this. One notable example is his research on Sweden. It used to be that Sweden created new noble titles and new noble houses, but that process ended in the 17th century. So if you have a noble surname in Sweden today, we know that your father's father's father's father's father's father's father (or whatever) was a member of the Swedish elite more than 300 years ago. By contrast, if you have the last name "Andersson" then that means that your great-great-great-great-great-great-great-great-great-grandfather wasn't a nobleman and probably didn't practice a skilled trade either. That's why he wound up with the generic surname. So we can look at the present-day incomes of people with noble surnames and compare them to the present-day incomes of people named "Andersson" and get a picture of the long-term persistence of the noble/Andersson class gap.
As you can see, there's quite a bit of persistence. And it's all the more striking precisely because this identification strategy is rather crude. A person with a noble surname could still be of mostly lower- or middle-class ancestry (my last name is Spanish because my paternal grandfather's father is Cuban, but the other 75 percent of my ancestry is Ashkenazi Jewish) and vice versa, so the surname thing should underestimate the long-term persistence of the class gap in Sweden. Clark's book, which will come out at the end of February, is apparently full of research along these lines.
*Correction, Jan. 23, 2014: A previous version of this post misspelled the first name of Emmanuel Saez.