Essays in Urban Economics

Preference heterogeneity is a driving force in the evolution of urban landscape. Combined with historical conditions, it can perpetuate existing inequality through residential sorting. This dissertation contributes to the literature in residential sorting and hedonic valuation to understand how preference heterogeneity affects location decisions and social welfare.

In chapter 1, I estimate the hedonic prices of particulate matter and nitrogen oxide using panel data from Glasgow, Scotland under a variety of functional form assumptions. I find that housing prices are the most elastic with respect to PM2.5 and the least elastic with respect to NOx. The hedonic price for all pollutants decreased from 2001 to 2011. At the median pollutant level, housing price elasticity of PM2.5, PM10, and NOx are -0.2 to -0.46, -0.17 to -0.48, and -0.05 to -0.3 respectively.

Chapter 2 examines the long term effects of neighborhoods on shaping individuals' religious attitudes and quantifies the implications for inequality. I develop and estimate a residential sorting model using new individual-level panel data in Glasgow, Scotland. I allow the marginal utility for neighborhood religious composition and marginal utility of income to vary by the individual's current income and childhood religious background. I find those with Catholic and non-Christian childhood backgrounds have much stronger religious homophily than those with Protestant/secular childhood background, and this religious homophily contribute to the intergenerational persistence of poverty traps.

Chapter 3 uses a household survey in Lahore, Pakistan to estimate a model of joint work and commuting mode choice accounting for heterogeneity in preferences and opportunity costs. Detailed information on individual trips across the city allows me to calculate the commute times for all neighborhood-to-local labor market combinations. I find that a nested logit model where individual are assumed to choose between work locations then choose commuting mode provides a good fit for the data. Estimates from the nested logit model suggest that travel time is a disamenity across all education levels, but less so as education level rises. The marginal willingness to pay (MWTP) to reduce commuting time by one minute per day ranges from 104 to 168 Rupees depending on education level.