Essays in applied econometrics
MetadataShow full item record
This thesis consists of three chapters. In chapter one a nested logit model is employed to capture workers' joint choice of residence location and mode of commuting in a U.S. metropolitan area. The nested logit model is estimated using tract level aggregated data from the U.S. Census Bureau's transportation planning package (CTPP) and summary file 3 from the 2000 U.S. decennial census. The estimation covers worker flows of more than four million work-residence census tract pairs contained in 275 metropolitan areas. The estimation is performed for four separate household income groups. The effects of accessibility to water bodies, limited access highways, central cities, and consumption opportunities in workers' decision process are considered. The robustness of the results is checked by estimating the joint choice model separately for a select number of large metropolitan areas representing different geographical parts of the U.S. The mode choice elasticity with respect to commuting time and residential location choice elasticity with respect to housing cost are found to agree reasonably well with existing estimates from other studies. In the 7 largest MSAs, there is a negative relationship between these two elasticities: MSAs with less public transit have a lower mode choice elasticity but a higher housing cost elasticity reflecting that when mode switches are more difficult, location adjustments may be more important. In chapter two a binary discrete choice model is used to analyze a profit maximizing developer's decision to construct a single family residence on a parcel of land. The role of expectations in the developer's choice decision is a major focus. Time series assessment data is used for parcels zoned single-family residential in Los Angeles County for twenty three years leading up to 2011. The estimates are used to discover the elasticity of the development probability and the housing supply with respect to housing price. The elasticity results are compared with housing supply elasticity estimates found in the extant literature all of which are based on different approaches and data. In Chapter 3 we use a large panel dataset covering the years 1988 to 2010 to estimate county specific total wage elasticities of labor demand for four highly aggregated industries in the United States. Our industries are construction, finance/real estate/service, manufacturing, and retail trade, which together employ on average over eighty percent of the U.S. national labor force per year. We use both the conventional constant coefficient panel data model and a random coefficients panel data model to estimate labor demand elasticities in various industries. We find the labor demand curves in all the industries studied to be downward sloping. We also find significant evidence that the total wage elasticity of labor demand exhibits regional variation. The labor demand estimates obtained in this study are useful to investigate the differential impact of various shocks and policy changes on the labor market. As an example, we use the estimated county specific labor demand elasticities to identify the impact of union membership and right to work laws on labor demand. We show that labor demand tends to become less elastic with higher union membership rates. We also find that labor demand becomes more elastic if a right to work law is in place.