Abstract
This dissertation consists of three essays relating to urban, transportation, and labor economics, all of which focus on challenges facing large cities. While the first and second chapters examine rising car use and migration in developing countries, the third chapter examines cities in California, fragmented by their size and traffic congestion. While the first chapter is a theoretical analysis and uses numerical simulations, the second and third chapters are empirical and use microdata on households and business establishments. Chapter 1, “Motorization in developing countries,” examines the rise in car use and decline in bus use in developing countries using a theoretical, mode choice model and numerical simulations. This analysis of commuter car/bus mode choice shows that in addition to rising income, other factors may drive rising car use at the urban level including: greater income inequality, which can both increase or decrease car use; traffic congestion, which hinders buses more than cars; and policy interventions, which can reduce congestion by maintaining bus service as an alternate travel mode, even as incomes rise. Chapter 2, “Migration and the next generation,” estimates the effect of migrating to a more developed region of a developing country on the educational attainment of migrants’ children by comparing migrants, who have moved from Brazil’s Northeast region to the more developed state of Sao Paulo, to non-migrants, who remain in the Northeast. Because migration is likely to be selective, this analysis uses state level instrumental variables of distance and past migration rates to identify the effect of migration. Instrumental variables estimation finds a negative effect, suggesting that migration may make children no better off, and possibly worse off. Chapter 3, “Access to workers and employers,” attributes economies of agglomeration to either labor market pooling or employer-based productivity spillovers by estimating the effect of access to same-industry employment, other-industry employment, and specialized workers using census tract level data for four industries. The results show that both access to specialized workers and access to same-industry employers contribute to economies of agglomeration and that the magnitude of the worker effect is large relative to employer-based productivity spillovers.