A Mobility Prophet

Only 26 years ago, one of the country’s most respected economists argued that social mobility was a problem largely solved. Birth rarely dictated a child’s destiny, suggested Gary Becker, in a 1987 speech to fellow economists. Children born rich often ended up poor, and children born poor often ended up rich. “Low earnings as well as high earnings,” said Mr. Becker, a University of Chicago professor who won a Nobel five years later, “are not strongly transmitted from fathers to sons.”

He had empirical evidence for his claim. Academic research at the time showed that the relationship between the earnings of different generations in a family was not especially strong. In technical terms, the correlation was so low: regardless of parents’ income, children’s income often reverted to the mean.

I first came across those findings, and Mr. Becker’s description of them, almost a decade ago, when reporting a series of articles for The Times on social class. At the time, Gary Solon, an economist of the University of Michigan, was one of the researchers considered to be on the cutting edge of mobility research. And Mr. Solon argued that Mr. Becker’s conclusions were mostly wrong.

Mr. Solon and other economists pointed out that tracking people over many years was quite difficult. The early studies of mobility, many of them done in the 1980s, suffered from these difficulties. As Janny Scott and I wrote in 2005:

Some studies relied on children’s fuzzy recollections of their parents’ income. Others compared single years of income, which fluctuate considerably. Still others misread the normal progress people make as they advance in their careers, like from young lawyer to senior partner, as social mobility.

When Mr. Solon and others looked at the new data that was emerging in the 1990s, they found that previous studies had confused statistical noise with evidence of a highly fluid society. When researchers were able to use accurate measures of people’s earnings, the relationship between the fortunes of parents and their children was quite strong. Most surprisingly, economists came to believe that a child’s chances of overcoming poverty in the United States were lower than in many other rich countries, despite our more egalitarian history.

“We all know stories of poor families in which the next generation did much better,” Mr. Solon, now a professor at Michigan State, told me almost a decade ago. “It isn’t that poor families have no chance.” But in the past, he added, “People would say, ‘Don’t worry about inequality. The offspring of the poor have chances as good as the chances of the offspring of the rich.’ Well, that’s not true. It’s not respectable in scholarly circles anymore to make that argument.”

At the time, even the new data was imperfect. Mr. Solon and other economists had to rely on surveys of several thousand families, some of whom dropped out of the panel, never to be surveyed again. (Mobility research, unlike research on inequality at a moment in time, depends on following the same people or families over many years.) In the last 10 years, however, the data on mobility has become better â€" much better. In a recent study, four economists were able to analyze millions of earnings records over more than three decades, as I reported on Monday.

One of the most striking aspects of this study, which economists say offers the most comprehensive picture of mobility yet, is how closely its findings match Mr. Solon’s. The numerical comparison of parents’ and children’s earnings â€" the statistical correlation between the two â€" is nearly identical, notes Jonathan Parker, a finance professor at M.I.T. (who did not work on the recent study).

Raj Chetty, a Harvard economist and one of the four co-authors, said the findings helped make clear Mr. Solon’s importance as a researcher. “What I find especially impressive is that many of his insights â€" most importantly that the U.S. has substantially lower mobility than previously thought â€" are basically borne out by our new data that is thousands of times larger,” Mr. Chetty said.

Sometimes, a huge batch of new information overturns old assumptions. But sometimes it confirms them.