I have to confess, I'm a little obsessed with measures of economic mobility. These measures address the basic elements of the American Dream. We get plenty of platitudes from politicians about the ability to start out from modest beginnings and achieve great success as an adult. But if you want to know the facts about opportunity, you have to look at economic mobility.
There is one major shortcoming of looking at economic mobility data, though. So far, most of the detailed measurements we've had on economic mobility in America only cover two generations: the people in their prime working years now, and their parents.
But time is on our side.
The University of Michigan's Panel Study of Income Dynamics (the longest running, nationally representative household survey in the world) has been running now for 44 years. And researchers are finally getting a chance to analyze the PSID's deep source of data across three generations of survey participants.
For the next two days, I'll be at a conference at the U of M where social scientists from across the world will present their findings from looking at the PSID and other data sources across multiple generations.
I've already had a chance to look at some of the papers that will be presented, and there are some tantalizing findings. Jean Yeung of the National University of Singapore and two co-authors from New York University looked at the black-white achievement gap across three generations. They found evidence that discrimination in the grandparent's generation had an impact on children's outcomes decades later.
Other papers look at the effects of extended family - aunts, uncles, cousins - to see how they affect economic mobility in other countries.
But I have to admit, the paper that's really caught my attention has nothing to do with the PSID. Economist Gregory Clark, from the University of California-Davis, examined mobility by tracking surnames, rather than tracking individual families. The results may be less precise than the measures of mobility we currently have, but the advantage is that Clark and his coauthors were able to consider many, many more generations.
Their working paper goes back as far as 1086 in England, not long after surnames came into existence. And tracking mobility like this across five, seven or even 10 generations of surnames comes up with some surprising results.
The working paper (a version of which can be found here) argues that the forces affecting mobility may have nothing to do with social policy, economic revolution or American exceptionalism.
Instead, it could all come down to a fixed mathematical law.
As Clark and his co-authors write in their working paper, "There seems to be a simple physics of social mobility."
I should emphasize that this paper has yet to be published, and the method for measuring mobility is very different from what's usually used among social scientists. Also, I'm a journalist, not a social scientist, so there may be problems that I'm missing. But still, the idea that mobility across generations is governed by one universal equation that doesn't change over time or across borders - that's a big idea. I can't wait to see how it's received at the conference.