Date of Degree


Document Type


Degree Name





Wim Vijverberg

Committee Members

Karna Basu

Chu-Ping Vijverberg

Subject Categories

Econometrics | Income Distribution | Labor Economics


social networks, group effects, spatial econometrics, New York City


Being a “social animal”, each person is inherently embedded into a complex structure of social relations. He has role models to aspire to, conformity rules to follow, and expectations to meet. This paper explores the different social influences each person experiences in life. Specifically, I consider how a person’s ethnic community, age reference group, occupational and industry group peers, and residential area neighbors affect his total income. I introduce a novel model of multiple social networks and discuss various identification implications. I apply the model empirically to New York City, which naturally is a very favorable environment to test for multiple social effects. The study uses generous person-level American Community Survey data that cover rich geographic, ethnic, demographic, and employment characteristics. The sample consists of five pooled cross-sections, resulting in about 270,000 observations. I analyze a model with spatial lags in dependent variable SARAR(1,0), a model with spatial lags in disturbance term SARAR(0,1), and a model with spatial lags in both dependent variable and error SARAR(1,1). I address self-selection into groups by correcting the empirical model for neighborhood choice and occupation choice. The main finding of the dissertation is that a person’s income is affected by his occupational group within his neighborhood, by people of the same occupation and industry, by people of the same age group, and by his ethnic group within his neighborhood. Additionally, people that work in the same industry and the same neighborhood have their unobserved characteristics correlated.