Urban Segregation: Implications for Productivity and Public Choice

Hannah Mead Kling

Advisor: Garett Jones, PhD, Department of Economics

Committee Members: Noel Johnson, Thomas Stratmann

Carow Hall, Conference Room
July 10, 2017, 11:00 AM to 01:00 PM


In this dissertation, I explore the causes and consequences of segregation through a public choice and urban economics framework. In Chapters 1 and 2, I examine whether municipalities adopt zoning regulations in order to exclude low-income and minority families. In the third chapter, I develop and calibrate a model of the productivity consequences of labor market segregation in Northern Ireland.

The first chapter, “Land-Use Regulations and Exclusion in Metropolitan Statistical Areas,” is co-authored with Garett Jones and Alex Tabarrok. In this chapter we analyze the public choice behind land-use regulation, specifically testing for an exclusionary motive, within metropolitan areas in the United States. We use an existing dataset of land-use regulation intensity among over 2,000 municipalities in the United States and Census demographic data. In general, we find support for a modest motive to exclude low-income households. A one-standard-deviation increase in the Gini coefficient is correlated with land-use regulations increased by approximately a tenth of a standard deviation. We find little to no support for a motive to exclude minority households. Income alone explains a large amount of variation in land-use regulations, with a one-standard-deviation increase in income correlating with approximately a third of a standard deviation higher land-use regulations. This finding is robust across specifications, and is consistent with motives other than exclusion for the adoption of land-use regulations.

In the second chapter, “Land-use Regulations As Exclusion: A GIS Analysis,” I extend the study of exclusionary land-use regulations to all municipalities in the contiguous United States. I use Geographic Information Systems (GIS) methods to identify the demographics within a set radius of each municipality for which land-use regulation data are available. If the desire to exclude low-income or minority households from a municipality drives the demand for land-use regulation, then larger demographic differences between a municipality and its surrounding areas would increase the incentive to adopt greater land-use regulations. I find evidence that greater income inequality in surrounding areas is correlated with higher land-use regulations. However, I do not find a robust correlation between racial or ethnic dissimilarity and land-use regulations. As a robustness check, I use a dataset of school district desegregation court orders as indicators of a municipality's exclusionary preferences. I find that these court orders are not robustly correlated with higher land-use regulation.

In the third chapter, “Modeling and Measuring Gains from Labor Market Desegregation in Northern Ireland,” I examine the impact of labor market segregation on productivity. Urban economic theories of agglomeration predict that productivity is higher in larger cities, due to better matching between employees and firms. This theory implies that in segregated economies, the gains from matching will be reduced. Over the past several decades, relations between Catholics and Protestants in Northern Ireland have gradually improved leading to decreased labor market segregation. I develop a model of the impact of labor force segregation on the agglomeration benefits of matching. Then, I use firm-level data to measure changes in employment segregation in Northern Ireland from 2001 to 2013. Finally, I use these estimates to calibrate the agglomeration model to estimate changes in output, wages, and number of firms from desegregation. The calibrated model estimates that a one percentage point decrease in segregation increases net output by 0.04 percent to 0.29 percent – approximately £5.3 million to £37.9 million.