Year of Publication


Degree Name

Doctor of Philosophy (PhD)

Document Type

Doctoral Dissertation


Business and Economics



First Advisor

Dr. James P. Ziliak


My dissertation investigates the reasons behind the persistence of income among individuals and US counties. I look at the role of initial conditions in explaining current level of income. In my first essay, I look at how childhood neighborhood conditions affect income of a person. To study persistence, I model income as an autoregressive process where the coefficient on the lagged dependent variable heterogeneous across individuals. In my second essay, I derive a new way to measure chronic poverty, or long term poverty. Current measures of chronic poverty cannot be used to compare improvements of poverty rates over time. Using my measure, one can compare to see if chronic poverty rates changed over time. My third essay looks at the historical reasons behind differences in income between rich and poor counties in the US. There are about 250 counties in the US where poverty rates have been above 20 percent for the last 40 years. I look at whether current and past factors, or differences in technologies is the main reason behind persistence of high rates of poverty in these counties.

Overall, I find that childhood neighborhood conditions have a big effect in determining the coefficient on the lagged dependent variable, that is, childhood neighborhood conditions affect persistence of income. I find that improving neighborhood poverty rates by one percentage point and father’s education by one year bring the greatest improvement of social welfare. In my second essay, I show the importance of measuring chronic poverty separately from total poverty; for example, between 2000 and 2005, total poverty declined, but chronic poverty rates actually increased, which shows that the long-term poor got worse off during that time period. In my last essay, I find that some US counties remained poor mainly because of differences in factor endowment, and past and present levels of human capital explain most of the differences in current level of income between poor and non-poor counties. Differences in factor endowments explained 80 percent of income between poor and non-poor counties, while technology accounted for only 20 percent of the difference.

Included in

Economics Commons