Carnegie Mellon University
2018_Navjeevan_Income_Mobility_in_America.pdf (2.48 MB)

Income Mobility in America

Download (2.48 MB)
posted on 2020-05-12, 19:13 authored by Manvendu Najeevan
This paper looks at both intragenerational and intergenerational income mobility in America. The paper is divided in two parts. The first part of the paper uses social security data to find that intragenerational income mobility in America has fallen by half over the past 50 years. I then consider which statistical model of income are consistent with the observed decline in income mobility. This paper finds that the AR(1) and ARIMA models do not do a good job of explaining mobility rates without using unreasonable parameters. However, the results from these models suggest that lowering the variance in shocks to income from year to year is more important in explaining falls in mobility than changes in year-to-year persistence of income. The model of income dynamics proposed by Guvenen (2016) does a better job of generating mobility rates seen in the data, but also provides somewhat limited explanation for falling mobility rates in our analysis. The results from the Guvenen model also suggest that variation in annual income shocks is significant in explaining falling mobility rates. In the second part I examine income mobility looking for significant county-level covariates to Chetty (2016) estimates on the causal effect of counties on income mobility. The analysis finds that social capital, income shares, and policy variables such as households on social security and size of public assistance in a county are significant in explaining variance in county level mobility rates. This paper also estimates the causal effects of housing rents as a percent of income using instrumental variable regression and estimates that a 1% increase in average housing rents as a percent of income in a county leads to a -0.21 change in the causal effect of spending an additional year in the county on percentile income rank at age 26. This paper also uses results from regressions to predict which counties would have the highest mobility rates in the future. Finally, the relationship between intergenerational mobility rates and inter-county migration is studied. The analysis finds that higher mobility rates are not strongly related to increased migration to a county, suggesting that movers have limited knowledge of which counties lead to better outcomes.





  • BS in Economics and Mathematical Sciences


Laurence Ales

Usage metrics

    Tepper Honors Theses


    Ref. manager