My research statement can be found here, It provides a fully summary of my PhD Thesis project. The following are short abstracts of my current research papers:

"What Matters for Choosing Your Neighbors? Evidence From Canadian Metropolitan Areas." (Job Market Paper, with Kristian Behrens). Submitted.


A corollary of the First Law of Geography and the Principle of Homophily is that “near things are more similar than distant things.” We test that proposition using spatially fine-grained data on thousands of ethnic colocation patterns in the six largest Canadian metropolitan areas. These colocation patterns reveal that groups that are more similar along various non-spatial dimensions—language, culture, religion, genetics, and historico-political relationships—colocate more. Our results provide a quantitative glimpse at the ‘deep roots’ of homophily and the role that different characteristics—such as language or religion—play in the ethnic stratification of cities. Our results hold for a wide range of alternative measures and different Census. We also find a heterogenous effect across cities, even an east-west gradient for some variables.

Presented at: National Housing Conference, November 21-22, 2018, Ottawa. 65th Annual North American Meetings of the Regional Science Association International, November 7-10, 2018, San Antonio, TX. 13th Meeting of the Urban Economics Association, October 12-13, 2018, Columbia university, New York City. 58th Congress of the Canadian Economic Association, Mai 9-11, 2018, Montreal.

  • "Distance-Based Segregation Measures." (with Kristian Behrens, preliminary draft available soon)


This paper provides new measures that help to overcome the weakness of existing measures of segregation, and allow a researcher to assess its magnitude. We apply point-pattern based measures of geographic concentration—used to assess the spatial clustering of plants—to the measurement of segregation. Our measures of excess segregation satisfy a number of desirable properties, encompass numerous existing approaches to segregation, and succeed in dealing with the problems raised in the literature. It allows us to finely assess the geographic distribution of different groups using spatially fine grained data and offer statistical testing of the observed patterns against various reference distributions, i.e., benchmarks. We show how our measures can be used to partly disentangle segregation by race from the geographic concentration of poverty. Our preliminary results show, for instance, the presence of segregation for black groups compared to a random benchmark. Moreover, after controlling for race, we found that poor black people are even more segregated than black people with higher incomes.

Presented at: 57th Congress of the Canadian Economic Association, Mai 10-12, 2017, Ottawa. 64th Annual North American Meetings of the Regional Science Association International, November 8-11, 2018, Vancouver, BC.

  • "Race, Poverty and Workplace: Testing the Spatial Skill Mismatch Hypothesis" (with Prottoy Aman Akbar)


We provide a measure for, and empirical application of, the physical disconnection between groups of people, belonging to some characteristics, and firms of certain sectors that are likely to hire these groups, a phenomenon known as Spatial Mismatch Hypothesis (SMH). On the individual side, we gathered 1990, 2000 and 2010 census data for New York Metropolitan Statistical Area (NYMSA), at the smallest geographic level, with information on race, education, sex and age. On the firm side, we gathered information on employment, sales, industry sector and the national distribution of skills by race. we aim to provide a continuous measure to: (1) assess spatial skill mismatch and show the physical disconnection (if there is any), (2) show that this physical disconnection have adverse labor outcomes, and (3) decompose the measure and see whether it is due to people’s or firm’s location.