This dissertation research examines immigration’s impact on metropolitan area’s economic growth.  It approaches this question with four dimensions: age effect, cohort effect, complement effect and legal status effect.  The research makes a major breakthrough in terms of methodology: it fully recognizes the spatial autocorrelation as a shortcoming of the traditional area-based immigration studies methodology and uses spatial regression to address this problem.  Also, the research contributes to the under-studied field of metropolitan level immigration and economic growth by composing metropolitan level micro-data on foreign-born population.   

Different from the previous immigration’s economic studies using classical microeconomic labor supply and demand model and focusing on wage and job opportunities, this research examines the macroeconomic measurement of regional economy: the regional Gross Domestic Product (GDP) and its growth rate.   The research observes immigration’s impact on regional economic growth through four effects, which correspond to four theoretical hypotheses: First, immigrant workers contribute to the regional economy disproportionally more because they bring a “population dividend” to the local economy - their population is more concentrated in the working age (age effect);  Second, immigrant workers contribute to the regional economy disproportionally more because they bring with them new ideas, rich international connections, exotic skills and the most recent cohort of arrival bears the largest value (cohort effect);  Third, immigrant workers contribute to the regional economy disproportionally more because they have the different skills to complement the native workers (complement effect); Fourth,  controlling for everything else, citizenship and legal permanent residency enable immigrant workers to contribute to the regional economy more greatly (legal status effect).    

Few immigration studies recognize the spatial autocorrelation of the immigration’s distribution. However, the fact that immigration is significantly spatially concentrated violates the basic assumptions of implementation of linear regression: a certain amount or quality of immigrants in one region does not necessarily have the same impact on the economic growth as it does in another region.  Some studies use a regional variable (e.g. South-east, Mid-west, East) to control the spatial auto-correlation.  This dissertation uses spatial econometrics methods (spatial regression and spatial panel regression) that more systematically address the spatial autocorrelation than the regional variable.   

In addition, because the limited data on immigration micro-data at the metropolitan level, it is rare that immigration is studied at the metropolitan level.  This research composes metropolitan level micro-data by pulling Public Used Microdata Area (PUMA) data provided in American Community Survey (ACS) into metropolitan statistical areas. Therefore, this dissertation makes important contribution to the metropolitan level immigration and economic growth research.