Syndicate content

Big Data Diving and US Intergenerational Income Mobility Hold Vital Lessons

Vamsee Kanchi's picture

Patterns of intergenerational income mobility in the United States reveal valuable lessons for economists and policy makers not just in this country, but also for the developing world, where successful efforts to promote shared prosperity and foster create better prospects for youth and children too often meet with frustration.

Raj Chetty, Professor of Economics at Harvard University lectured on this topic recently at the World Bank. For his talk, Chetty drew on recent research by him, Nathaniel Hendren, Patrick Kline and Emmanuel Saez. Chetty and team analyzed anonymous tax records on earnings of 40 million US children and their parents to gauge a child's chances of moving up the income distribution relative to his or her parents.

America is often viewed as the land of opportunity, but Chetty finds that the income gap between an American born in a lower-income family at the bottom 25th percentile, and another born in the top 25 percentile when both are 30 years old is 34 percent.  In countries like Denmark, however, children from lower income families get closer to the median income faster than those in the US - the gap is 18 percent.

While cross-country comparisons of income mobility are important, the bulk of Chetty’s work focuses on income mobility within the US.

On average, a child starting off at the 25th percentile in Salt Lake City, Utah will end up at the 46th percentile at age 30, but another from Charlotte, NC will end up only at the 36th percentile, earning $21,000 a year.  Where a child grows up is correlated with things like his or her likelihood of attending college, or having a child as a teenager; also, children benefit proportionally in relation to how early in their life their parents move to a region with greater upward mobility.

Stronger racial segregation, commuting time, levels of inequality in a given region, social capital (such as church attendance), family structure and school quality are also correlated with intergenerational income mobility.

To put it bluntly: it really matters where a child from a low income family in the US lives.

The New York Times has done a nice job of visually presenting Chetty’s findings, and you can find maps and charts here.  Also see a new article by David Leonhardt here.

More details of Chetty and team’s findings are also available from the video of his lecture, and the related Website.

Of the correlated factors mentioned above, Chetty focuses on improving quality of education - particularly the quality of teachers - as a potential fix for immobility, mostly because academics know how to change education, and to document that change. However, the other factors mentioned above should also be investigated and considered when it comes to policy discussions.

But what does this mean for readers of this blog, who are primarily focused on concerns of the developing world?

In line with Chetty’s views on education, research on India finds that improving teacher attendance leads to bigger impact, and that paying teachers for performance increases test scores.

Chetty and team also think there is room when tackling economic development issues to focus on place-based policies. Government officials (think city mayors) are already doing this kind of thing by targeting interventions to specific areas within their constituencies.

But perhaps the scope for the biggest impact is in using Big Data to inform policymaking.  In Chetty’s work, focusing on the US and Scandinavian countries provided deep insights - collecting anonymized data on 40 million individuals would have been impossible only a few years ago - but his work would have benefited by comparing data from more countries.

It is also thanks to Big Data that he was able to conduct quasi-experiments to assess the impact of teacher quality.  By looking at school district records of 2.5 million children, and their tax records some 20 years later, Chetty and team were able to identify correlations between teacher quality and the effect on income earnings, as well as college attendance and teenage pregnancies.

Given the challenges in collecting large amounts of data (not the least of which is concern for privacy and citizen rights), there clearly are difficulties involved, but the positive impacts, as demonstrated through Chetty’s work in the US, can be significant.

Add new comment